Startseite Quantitative and qualitative interpretation of community partitions by map overlaying and calculating the distribution of related geographical features
Artikel Open Access

Quantitative and qualitative interpretation of community partitions by map overlaying and calculating the distribution of related geographical features

  • Zhang Haitao EMAIL logo , Ji Kang , Shen Huixian , Song Rui , Liu Jin Yuan und Yang Yu Xin
Veröffentlicht/Copyright: 7. November 2023
Artikel downloaden (PDF)
Veröffentlichen auch Sie bei De Gruyter Brill
Manuskript einreichen Informationen für Autor*innen
Open Geosciences
Aus der Zeitschrift Open Geosciences Band 15 Heft 1

Abstract

Applying community detection algorithms in spatial interaction networks constructed from modern human communication records is an essential means of evaluating urban territorial subdivisions. Previous studies have usually involved qualitative rather than quantitative interpretations of community detection results. This article proposes a method of quantitatively and qualitatively interpreting community partition results by map overlaying the spatial regions corresponding to the detected communities with the related geographical features and by calculating the distribution of the geographical features contained in the regions and the entropy value of each distribution. The interpretation of the communities detected from the spatial interaction networks is carried out from the perspective of multi-temporal and multi-spatial scales and multi-geographical features. Extensive experiments were conducted with Milan, Italy, as the study area. The spatial interaction records reflected by telephone calls, land use, and point of interest (POI) data were used as the experimental data. Experimental results demonstrated the effectiveness of our method, and the specific results include: (1) Qualitative interpretation of multi-spatial resolution scale communities detected from the long-term aggregated spatial interaction network. The cohesiveness, homogeneity, and heterogeneity of the detected communities were qualitatively interpreted by the spatial distribution patterns of the land use dataset and the POI dataset. (2) Quantitative interpretation of multi-spatial resolution scale communities detected from the long-term aggregated spatial interaction network. The low spatial resolution scale community partitions and the high spatial resolution scale community partitions were interpreted through the statistical distribution of the land use dataset and the POI dataset, respectively. (3) Qualitative interpretation of the stable and active regions discovered from the community time series. Regardless of the community partitions’ spatial resolution scales, the stable and active regions were distinguished with the statistical distributions of the land use dataset and the POI dataset.

Keywords: spatial interaction network; community partitions; geographical features; statistical distribution

1 Introduction

In recent years, modern human communication has undergone massive structural changes in the past few decades, which have produced databases of personal communication data, such as mobile phone call records. Personal communication data can reflect the interactions between spatial regions with a high spatial resolution. Spatial interactions naturally form networks/graphs, often referred to as spatial interaction networks, in which each node is a location and each link is an interaction between two nodes [1]. Community detection algorithms from spatial interaction networks have been proposed in many studies to evaluate the existing administrative divisions of urban [2].

Previous studies were qualitative rather than quantitative in their interpretation of the community partition results. For example, several studies either reported the experimental results (i.e., significant overlaps occurred between the community’s borders and the existing administrative borders) or described frequent correlations between community partitions and cultural, linguistic, or topographical features, etc. However, quantitative analysis of community partitions plays a vital role in accurately understanding the socioeconomic factors of the territorial subdivision results. It even provides decision-making assistance for urban planning, traffic management, etc. Specifically, for urban planning departments, accurately obtaining the number and types of geographic elements contained in each community is an essential factor in evaluating the rationality of planning effects.

Therefore, we proposed a method to interpret the communities detected from spatial interaction networks constructed from mobile phone call records quantitatively and qualitatively by map overlaying and calculating the distribution of related geographic elements in the communities. In the rest of this article, we reviewed the relevant related works in Section 2 and described the study area and data in Section 3. Section 4 presented the proposed methodology. Experimental results are discussed in Section 5, followed by conclusions in Section 6.

2 Related work

The problem of community detection has been studied extensively in recent years. With the increasing availability of large network datasets, fast community detection algorithms have been focused on in recent years [3]. The leading approach is based on optimizing the network’s modularity, which was introduced by Newman [4]. Furthermore, inspired by the fact that in most large networks, there are several natural organization levels by which communities divide themselves into sub-communities [5]. Blondel et al. proposed a community detection method that reveals this hierarchical structure [3], known as the Louvain method. The Louvain method provides the optimal number of groups and their composition automatically, with no outside intervention; that is, the number of groups and the composition of the groups are endogenous [6]. In addition, the Louvain method is well-defined mathematically, has been used successfully in many contexts, and is available in several software packages for the automatic processing of networks, such as the igraph library (http://igraph.sourceforge.net/) and NetworkX (http://networkx.lanl.gov/) [5].

Scholars have studied community detection methods utilizing personal communication data, which mainly use the flow of information between spatial regions to reflect spatial interactions. The classification for these methods is based on technique criteria: (1) applying only the topological structure and (2) applying both the topological structure and the spatial structure. Examples of the typical former methods include the following. Blondel et al. [3] proposed a heuristic method based on modularity optimization and extracted the community structure of the Belgian mobile phone network of 2.6 million customers. Ratti et al. [7] adopted a modularity function as the objective function to delineate borders emerging from a network extracted from a large database of telecommunication records in Great Britain. Sobolevsky et al. [2] used an extended set of countries and clustering indices to quantify the overlaps of borders. They provided ample additional evidence for their observations using phone data from countries of various scales across Europe, Asia, and Africa (i.e., France, the UK, Italy, Belgium, Portugal, Saudi Arabia, and the Ivory Coast). Kallus et al. [8] developed a network community analysis based on one of the most popular online social networks (i.e., Twitter), which represents the ties between more than 5.8 million geo-located users in North and South America, Europe, and Asia. In addition, Gao et al. [9] proposed a modularity function incorporating a gravity model to determine the clustering structures of spatial-interaction communities of call flow and movement flow from a mobile phone dataset containing 1 week of calls in a city in China. Botta et al. [10] presented a detailed analysis of the community structure of the network of mobile phone calls in the metropolitan area of Milan, revealing temporal patterns of communications between people. Xu et al. [11] analyzed a mobile positioning dataset that captured the nationality and movement patterns of foreign tourists to South Korea. They revealed the collective dynamics of tourist movements and critical differences across nationalities. Wang et al. [12] developed a methodological framework for detecting urban mobility structure at the transportation analysis zone level in Beijing using mobile phone signal data. They identified the hidden structure of urban mobility extracted from phone data.

The approaches described earlier applied the topological structure of a network to determine community partitions, but few considered the spatial character of the network. Many spatial interaction networks are spatially constrained in geographical space. That is, nearer objects are more related than distant objects; however, the relationships between vertices should be defined by not only the links connecting them but also the distance between them. Chen et al. [13] proposed a geo-distance-based method in a spatially constrained network to identify communities that are both highly topologically connected and spatially clustered. Koylu [14] used a spatially constrained hierarchical regionalization algorithm to reveal multi-scale community structures within the interpersonal communication network on Twitter.

3 Study area and data

3.1 Study area

This study was conducted in Milan, Italy. As the second-largest city in Italy, Milan is also an important city in Europe. The telecommunications company in Milan publicly provided mobile phone data, and Copernicus provided us with geographic feature data for the area. Then, we chose an urban region of Milan as the study area, which is shown in Figure 1.

Figure 1 Study area.
Figure 1

Study area.

3.2 Telephone dataset

The telephone dataset was provided by the first edition of the Big Data Challenge launched by Telecom Italia (https://pan.baidu.com/s/1wF7Z_zjQcr2lCKwdqYn08w, Extraction code:3scu). The telephone dataset was collected for one and a half months (between 1 November 2013 and 14 December 2013). It was spatially aggregated into a non-overlapping spatial tessellation of 10,000 grids, each with dimensions of 235 m by 235 m. The overlay map of the grids and the study area is shown in Figure 2.

Figure 2 The overlay map of the study area overlaid by 10,000 grids.
Figure 2

The overlay map of the study area overlaid by 10,000 grids.

Furthermore, the telephone dataset was aggregated in time slots of 10 min. Finally, for each pair of spatial grids, a spatial interaction record was obtained within 10 min. Examples of spatial interaction records are given in Table 1.

Table 1

Examples of spatial interaction records

Square id1 Square id2 Time interval Directional interaction strength
1066 128 1,388,561,400,000 1.1468710926626244 × 10−4
1066 1281 1,388,579,400,000 7.573166813008205 × 10−4
1066 1282 1,388,595,000,000 9.882422919705961 × 10−5
… … … …

In Table 1, Square id1 and Square id2 are the ids of the origin grid and the destination grid of a spatial interaction, respectively. The time interval is the time slot of 10 min. The Directional interaction strength is the directional interaction strength between Square id1 and Square id2, which is proportional to the number of calls exchanged between callers located in Square id1 and receivers located in Square id2. In total, the telephone dataset was aggregated into 6,404,487,297 spatial interaction records between 10,000 grids.

3.3 Geographical features dataset

The land use dataset and the point of interest (POI) dataset are two types of typical feature datasets that are closely related to human activities. The land use dataset has a long period and belongs to the administrative planning data, which can reflect the long-term activity law of the population. The POI dataset has a short period and can have higher spatial and temporal accuracy than the land use dataset, which can reflect the short-term laws of human activities, and has a positive effect on capturing daily crowd activities. Therefore, in this article, we choose these two kinds of typical geographic feature datasets as the experimental evaluation data for community division results. The land use dataset was collected by Earth observation satellites and was combined with observation data from sensor networks on the Earth’s surface collected in 2012. These data were provided by Copernicus, which is a European program for monitoring the Earth (see https://land.copernicus.eu/local/urban-atlas/urban-atlas-2012?tab=download). The land use dataset includes 21 land use types, for which the overlay map with the grids is shown in Figure 3.

Figure 3 The overlay map of the land use dataset with the 10,000 grids.
Figure 3

The overlay map of the land use dataset with the 10,000 grids.

The POI dataset was derived from the Open Street Map (https://www.openstreetmap.org/). The data are classified into eight categories: transportation services, recreation, business, public services, catering and accommodations, party and government organizations, sightseeing, and shopping. The overlay map for the POI dataset and the grids is shown in Figure 4.

Figure 4 The overlay map of the POI dataset overlaid by the 10,000 grids.
Figure 4

The overlay map of the POI dataset overlaid by the 10,000 grids.

4 Research methods

The proposed method includes three phases: spatial interaction network extraction, community detection, and geographical feature overlay. The steps include: (1) extracting the network from the telephone interaction data, with the requester and the receiver as nodes, and their interaction records as edges and weights, thus constructing a huge telephone interaction network; (2) applying the community detection algorithm to achieve different levels of community structure in the telephone interactive network; and (3) overlaying geographic features on the community structure and using qualitative and quantitative methods to interpret the results of community division.

4.1 Spatial interaction network extraction

A spatial interaction network was constructed from the spatial interaction records aggregated from the telephone dataset. A node in the network represents a grid, and an edge encodes a call flow between a pair of grids. The edge was directed and weighted, and the weight was proportional to the directional interaction strength between the origin node and the destination node of the edge. A spatial interaction network extraction involves two primary definitions.

Definition 1

Given a non-overlapping spatial tessellation Grids = { g 1 , g 2 , … , g n } , where g i ,   1 ≤ i ≤ n represents a grid, a spatial interaction record is defined as SI = ( ori_grid , des_grid , interval , intensity ) , where ori_grid and des_grid ∈ Grids represent the origin grid and the destination grid of the spatial interaction record, respectively, interval represents the temporal aggregation interval of the telephone dataset between the two grids, and intensity represents the interaction strength between the two grids. The example corresponding to this definition can be found in Figure A1 in the Appendix.

Definition 2

For a set of spatial interaction records SIs = { si 1 , si 2 , … , si n } and a non-overlapping spatial tessellation Grids = { g 1 , g 2 , … , g m } , SIN = { V , E } is defined as a spatial interaction network, where V = { v 1 , v 2 , … , v s } , v i ·grid ∈ Grids , and 1 ≤ i ≤ s . That is, for the node v i ∈ V , there is the corresponding grid v i ·grid ∈ Grids ; E = { e 1 , e 2 , … , e t } , 1 ≤ t ≤ ( m × ( m − 1 ) ) ; and the condition that for any directed edge e = { v o , v p } ∈ E is satisfied. In addition, there is a spatial interaction record si ∈ SIs , si· ori_grid = v o · grid , si·des_grid = v p · grid and si · intensity = e · intensity . The example corresponding to this definition can be found in Figure A2 in the Appendix.

The number of experimental spatial interaction records is usually very huge, and the data used in the experiments in this study include about 6.4 billion records. Therefore, to overcome the computational difficulties associated with using large datasets, the spatial interaction network generation algorithm can be implemented in a big data computing platform, such as the Apache Spark graph parallel computing engine GraphX. The implementation algorithm is as follows.

Algorithm 1

NetworkGen (SIFile, ref Graph)

Input: SIFile represents the spatial interaction records file.

Output: Graph represents the generated spatial interaction network.

1. $/path/to/spark/bin/spark-shell

2. import org.apache.spark.graphx.

3. val sparkConf = new SparkConf ().setAppName(“”).setMaster (“”)

4. val sc = new SparkContext(Conf)

5. val SIdata = sc.textFile (SIFile)

6. val edges:RDD[Edge[Int]]=patterndata map(

7. line= >

8. val row = line split” “

9. Edge(row(0).toInt,row(1). toInt, row(2). toDouble)

10. val graph: Graph[Int,Int] = Graph.fromEdges(edges, 1)

11. val uniqueInputGraph = egograph.groupEdges((e1, e2) = > e1 + e2)

Line 1 starts sparkshell; Line 2 imports the GraphX library; Lines 3–5 load the spatial interaction records file; Lines 6–9 individually scan the first, second, and third string in the records file as the origin, destination nodes, and intensity of an edge, respectively. Line 10 generates the spatial interaction network from the edges. Line 11 groups the repeated edges in the network. That is, it assigns weights to the edges.

4.2 Community detection

Community detection using a spatial interaction network can be formally defined as follows.

Definition 3

Given a spatial interaction network SIN = { V , E } , SINC = { V ′ , E ′ } is defined as a community detected by SIN , where V ′ ⊆ V , E ′ ⊆ E , V ′ ≠ ∅ , and E ′ ≠ ∅ . The example corresponding to this definition can be found in Figure A3 in the Appendix.

Since our telephone interaction data are not high-dimensional, the current emerging deep-learning methods are not suitable for our research. The Louvain algorithm uses a two-layer iterative procedure, which can avoid the disadvantage that the nodes cannot be separated after merging the nodes in the simple agglomeration method. Therefore, we chose the Louvain algorithm for our study. The implementation algorithm is as follows.

Algorithm 2

CommunityGen (Graph, Community)

Input: Graph represents the spatial interaction network.

Output: Community represents the generated community.

  1. var louvainGraph = createLouvainGraph(graph)

  2. var level = −1

  3. var q1 = −1.0

  4. var halt = false

  5. var minProgress = 1

  6. var progressCounter = 1

  7. do {

  8.    level+ = 1

  9.    val (currentQ, currentGraph, passes) = louvain(sc, louvainGraph,

  10.    minProgress, progressCounter)

  11.    louvainGraph.unpersistVertices(blocking = false)

  12.    louvainGraph = currentGraph

  13.    saveLevel(sc, level, currentQ, louvainGraph)

  14.    if (passes > 2 && currentQ > q1 + 0.001) {

  15.       q1 = currentQ

  16.       louvainGraph = compressGraph(louvainGraph)

  17.      }else { halt = true }

  18.   } while (!halt)

Line 1 converts the input network graph into a Louvain graph structure that is convenient for realizing Louvain’s algorithm. Lines 2–6 initialize the algorithm parameters. Lines 7–17 describe the process of generating hierarchical communities, which will produce three levels of communities in total. The definition of modularity current Q is extended to the scenario of weighted directed networks in this article, as follows:

Q = 1 2 w ∑ i ∑ j w ij − w i out w j in 2 w δ ( C i , C j ) ,

where w i out = ∑ j w ij and w j in = ∑ i w ij are, respectively, the output and input strengths of nodes i and j , and the total strength is 2 w = ∑ i ∑ j w ij . δ ( C i , C j ) is the Kronecker delta function, which takes the value 1 if nodes i and j are in the same community and 0 otherwise [15].

Communities are usually time-dependent and evolve. The concepts of a stable region and an active region can be defined according to the spatial range change of the community in the two time periods. The two concepts can be formally defined as follows.

Definition 4

Given two communities SINC A and SINC B , which are, respectively, detected from two spatial interaction networks SIN A and SIN B at different periods, Ints V ( A , B ) = { V } is defined as the intersection of SINC A and SINC B nodes, where for any node v ∈ V , the condition is satisfied: v · grid ∈ ( SINC A · Grids ∧ SINC B · Grids ) .

Furthermore, the condition is satisfied: ( Ints V ( A , B ) · Grids · count ≥ SINC A · Grids · count / 2 ) ∧ ( Ints V ( A , B ) · Grids · count ≥ SINC B · Grids · count / 2 ) , Ints V ( A , B ) · Grids is defined as the stable region corresponding to SINC A and SINC B , and if ( ( Ints V ( A , B ) · Grids − SINC A · Grids ) ≠ ∅ ) ∨ ( ( Ints V ( A , B ) · Grids − SINC B · Grids ) ≠ ∅ ) , then ( Ints V ( A , B ) · Grids − SINC A · Grids ) , or ( Ints V ( A , B ) · Grids − SINC B · Grids ) is defined as the active region corresponding to SINC A and SINC B . An example can be found in Figure A4 in the Appendix.

4.3 Geographical feature overlay

The overlay operation and statistical analysis of the spatial data are essential functions of geographic information system (GIS). Depending on the geometric type of the spatial data, there will be different implementation methods. In this study, we used two GIS functions: the polygon overlay and point-in polygon functions. Specifically, we used the polygon overlay function to sum the land use areas intersected by the grids corresponding to the community. We also used the point in polygon function to count the number of POIs in the grids corresponding to the community. This geographical feature overlay process involves six basic definitions.

Definition 5

For a set of land use types LandTypes = { lt 1 , lt 2 , … , lt n } , LandData = { ld 1 , ld 2 , … , ld m } is defined as a land use dataset, where ld i = ( geo , lt ) ,   0 ≤ i ≤ m , lt ∈ LandTypes represents a plot of land, and geo represents the region where the plot is located. The example corresponding to this definition can be found in Figure A5 in the Appendix.

Definition 6

Given a land use dataset LandData = { ld 1 , ld 2 , … , ld n } and a community SINC = { V , E } , V = { v 1 , v 2 , … , v m } , the overlay operation between SINC and LandData can be defined as SIN C Lan d OL = { ol 1 , ol 2 , … , ol m } , where ol i = { v i , polygon overlay ( v i · grid , ld 1 . geo ) , polygon overlay ( v i · grid , ld 2 . geo ) , … , polygon_overlay ( v i · grid , ld n . geo ) } ,   1 ≤ i ≤ m . If the topological intersect relationship is satisfied, the polygon_overlay function will return the area that v i · grid intersects ld j . geo ,   1 ≤ j ≤ n , and the land use type ld j . lt . The example corresponding to this definition can be found in Figure A6 in the Appendix.

Definition 7

Given a land use type set LandTypes = { lt 1 , lt 2 , … , lt n } , a land use dataset LandData = { ld 1 , ld 2 , ld 3 , … , ld m } , a set of communities SINC s = { sinc 1 , sinc 2 , sinc 3 , … , sinc s } , and the overlay set between SINC s and LandData : SINC_Land_OLs = { sl ol 1 , sl ol 2 , slol 3 , … , sl ol t } , the quantitative statistical analysis of SINC_Land_OLs can be defined as follows:

SIN C Lan d Sta = { ( c i , lt j , are a ratio ) } ,   1 ≤ i ≤ s ,   1 ≤ j ≤ n ,

where c i ∈ SINCs represents the community number, lt j ∈ LandTypes represents the land use type, and are a ratio = area ( lt j ) area ( c i ) represents the ratio of the area of the land use type lt j to the area of community c i . The example corresponding to this definition can be found in Figure A6 in the Appendix.

Similarly, we can obtain the definitions of the POI data overlay with the grids corresponding to the community.

Definition 8

For a set of POI types POITypes = { pt 1 , pt 2 , … , pt n } , POIdata = { pd 1 , pd 2 , … , pd m } is defined as a POI dataset, where pd i = ( geo , pt ) ,   0 ≤ i ≤ m , pt ∈ POITypes represents a POI type, and geo represents the point where the POI is located. The example corresponding to this definition can be found in Figure A7 in the Appendix.

Definition 9

Given a POI dataset POIData = { pd 1 , pd 2 , … , pd n } and a community SINC = { V ′ , E ′ } , V = { v 1 , v 2 , … , v m } , the overlay operation between SINC and POIData can be defined as SINC_POI_OL = { ol 1 , ol 2 , … , ol m } , where ol i = { v i , point_in_polygon ( v i · grid , ld 1 . geo ) , point_in_polygon ( v i · grid , ld 2 . geo ) , … , point_in_polygon ( v i · grid , ld n . geo ) } ,   1 ≤ i ≤ m . If the point-in polygon topological relationship is satisfied, the point_in_polygon function will return the POI type ld j . lt ,   1 ≤ j ≤ n . The example corresponding to this definition can be found in Figure A8 in the Appendix.

Definition 10

Given a set of POI types POITypes = { pt 1 , pt 2 , … , pt n } , POIdata = { pd 1 , pd 2 , … , pd m } , a set of communities SINCs = { sinc 1 , sinc 2 , sinc 3 , … , sinc s } , and the overlay set between SINCs and POIData : SINC_POI_OLs = { sl ol 1 , sl ol 2 , slol 3 , … , sl ol t } , the quantitative statistical analysis of SINC_POI_OLs can be defined as SIN C POI Sta = { ( c i , pt j , count ratio ) } ,   1 ≤ i ≤ s ,   1 ≤ j ≤ n , where c i ∈ SINCs represents the community number; pt j ∈ POITypes represents the POI type; and count ratio = count ( pt j ) count ( c i ) represents the ratio of the count of the POI type pt j to the count of community c i . The example corresponding to this definition can be found in Figure A8 in the Appendix.

In information theory, the entropy of a random variable is the average level of “information,” “surprise,” or “uncertainty” inherent to the variable’s possible outcomes [16]. Entropy is one of the several ways to measure biodiversity and is applied in the form of the Shannon index [17]. In this article, we use information entropy to measure the distribution patterns of features in various communities.

Therefore, for the statistical analysis of SIN C Lan d Sta , we can obtain the definition of the entropy of the land use distribution in the community as follows:

Definition 11

For a quantitative statistical analysis result SIN C Lan d Sta = { ( c i , lt j , are a ratio j ) } ,   1 ≤ i ≤ s ,   1 ≤ j ≤ n , the information entropy of SIN C Lan d Sta can be defined as SIN C entropy landuse = { ( c i , entropy i ) } ,   1 ≤ i ≤ s , where entropy i = − ∑ 1 n area ratio j × ln area ratio j . The example corresponding to this definition can be found in Figure A6 in the Appendix.

Furthermore, for the statistical analysis of SIN C POI Sta , we can obtain the definition of the entropy of the community and the POI distribution as follows:

Definition 12

For a quantitative statistical analysis result SIN C POI Sta = { ( c i , pt j , count ratio j ) } ,   1 ≤ i ≤ s ,   1 ≤ j ≤ n , the information entropy of SIN C POI Sta can be defined as SIN C entropy = { ( c i , entropy i ) } ,   1 ≤ i ≤ s , where entropy i = − ∑ 1 n count ratio j × ln count ratio j . The example corresponding to this definition can be found in Figure A8 in the Appendix.

5 Results and analysis

The experiments were performed on an Apache Spark standalone cluster computing framework (v2.2.0). The cluster uses a master/worker architecture and is deployed in three executors, a master and two workers, each of which has eight cores (Intel Xeon E5-2609 v3 CPU with a clock speed of 1.90 GHz) and 72 GB RAM [18].

To improve the computational efficiency of our proposed method, we further conducted temporal aggregation of the spatial interaction data aggregated at a time granularity of 10 min. According to Definitions 1 and 2, for the spatial interaction records from November 1 to 14, 2013, one aggregation network was constructed, which included 10,000 nodes and 1,169,475,402 edges; For the spatial interaction records from November 15 to December 14, 2013, 29 aggregation networks were constructed, each containing 1 day, and the basic information for the 29 networks is presented in Table 2.

Table 2

The basic information for the 29 networks

Aggregation time Node count Edge count Aggregation time Node count Edge count
2013-11-15 10,000 89,739,230 2013-12-01 9,998 48,113,969
2013-11-16 10,000 64,408,794 2013-12-02 9,998 80,680,862
2013-11-17 10,000 49,249,503 2013-12-03 9,998 82,257,246
2013-11-18 10,000 85,476,820 2013-12-04 9,998 82,309,164
2013-11-19 10,000 86,342,834 2013-12-05 9,998 81,894,749
2013-11-20 10,000 85,462,864 2013-12-06 9,998 84,313,710
2013-11-21 10,000 85,383,898 2013-12-07 9,998 57,065,533
2013-11-22 10,000 84,626,348 2013-12-08 9,998 46,498,383
2013-11-23 10,000 59,987,375 2013-12-09 9,998 81,602,904
2013-11-25 9,998 81,136,059 2013-12-10 9,998 82,930,993
2013-11-26 9,998 81,861,672 2013-12-11 9,998 85,054,332
2013-11-27 9,998 80,019,399 2013-12-12 9,998 84,677,570
2013-11-28 9,998 80,860,167 2013-12-13 9,998 86,467,351
2013-11-29 9,998 82,408,228 2013-12-14 9,998 62,433,177
2013-11-30 9,998 60,682,454

In Table 2, Aggregation time represents the date we collected data and aggregated it. Node count is the number of all nodes in the spatial interaction records aggregated. Edge count is the number of all edges in the spatial interaction records aggregated.

5.1 Qualitative and quantitative interpretation of the communities detected from a long-term aggregated spatial interaction network

The qualitative and quantitative interpretation of the communities detected from the spatial interaction networks is carried out from the perspective of multi-temporal and multi-spatial scales and multi-geographical features. Specifically, multi-temporal scales refer to interpreting the communities detected from long- and short-term aggregated spatial interaction networks. Multi-spatial scales refer to the interpretation of communities detected from spatial interaction networks at multiple spatial resolution levels. Multi-geographical features refer to the quantitative and qualitative interpretation of the stable regions and active regions found in the community time series using two typical geographical features (i.e., land use type and POI).

5.1.1 Multi-spatial resolution scale communities detected from a long-term aggregated spatial interaction network

We utilized the Louvain method to detect the communities from the network constructed from the aggregated half-month spatial interaction records. According to Definition 3, we use Louvain’s method to detect the community. The number of communities and the size of communities are essential parameters, but it is not necessary to set these two parameters for the Louvain algorithm. By adopting an iterative bottom-up approach and only three passes, we discovered three-level hierarchical community structures without having to fix the number of communities or their sizes, which are shown in Figure 6. The colors in Figure 6 are of no particular significance and are intended to facilitate easier reading of the map. In detail, during the first pass, the 10,000 nodes of the network were partitioned into 32 communities, which consist of the first level of the hierarchical community structures, hereafter, referred to as high spatial resolution community. After the second pass, the 32 intermediate nodes were partitioned into eight communities, which consist of the second level of the hierarchical community structures, referred to as medium spatial resolution community. During the third pass, three communities remained, which consisted of the last level of the hierarchical community structures, referred to as low spatial resolution communities.

As seen in Figure 5, the three-level hierarchical community structures yield geographically cohesive regions, even though we did not impose constraints regarding the proximity or contiguity of the grids in the communities. The results of this experiment demonstrated that the spatial interaction reflected by the aggregated telephone call records indicates the underlying dependence of social bonding on distance. That is, people still most often call those who are geographically close and whom they probably see often, which is consistent with the findings of previous studies [6].

Figure 5 Multi-spatial scale communities detected from a long-term spatial interaction network. (a) Low spatial resolution community partitions. (b) Medium spatial resolution community partitions. (c) High spatial resolution community partitions.
Figure 5

Multi-spatial scale communities detected from a long-term spatial interaction network. (a) Low spatial resolution community partitions. (b) Medium spatial resolution community partitions. (c) High spatial resolution community partitions.

Furthermore, from Figure 5, we also found that the geographic spatial region of community number 1 in the low spatial resolution community partitions (Figure 5(a)) does not change in the medium spatial resolution community partitions (Figure 5(b)) or in the high spatial resolution community partitions (Figure 5(c)). In contrast, the other communities combine with their neighboring communities to generate communities with more significant geographically cohesive regions. That is, some communities are homogeneous on multilevel spatial scales, while other communities are different. Specifically, the communities are homogeneous on low spatial resolution scales but are heterogeneous on high spatial resolution scales.

5.1.2 Qualitative interpretation of communities through spatial visualization of map overlay with geographical features

We qualitatively interpreted the cohesiveness, homogeneity, and heterogeneity by spatial visualizing of map overlay communities with geographical features. According to Definitions 5 and 6, we first map overlaid the land use dataset with the spatial regions corresponding to the three-level hierarchical community partitions as shown in Figure 6. The results of the low-, medium-, and high spatial resolution community partitions are shown in Figure 6(a)–(c), respectively.

Figure 6 The overlay map of the three-level hierarchical community partitions with the land use dataset. (a) Low spatial resolution. (b) Medium spatial resolution. (c) High spatial resolution.
Figure 6

The overlay map of the three-level hierarchical community partitions with the land use dataset. (a) Low spatial resolution. (b) Medium spatial resolution. (c) High spatial resolution.

As seen in Figure 6(a), the spatial distribution patterns of the land use dataset contained in the corresponding spatial regions of the three communities are entirely different. Moreover, as seen in Figure 6(b), the spatial distribution patterns of the land use dataset contained in the corresponding spatial regions of the eight communities can also be distinguished. We can intuitively interpret the cohesiveness and heterogeneity of the detected communities through graphical visualization. Specifically, for community 1 in Figure 6(a), the reason for its homogeneity in all three levels of the hierarchical community partitions is that the spatial distribution pattern of the land use dataset in the community has not changed at all. Moreover, for communities 2 and 3 in Figure 6(a), the reason for their heterogeneity in the three-level hierarchical community partitions is as follows. In Figure 6(b), communities 3 and 2 in Figure 6(a) are divided into sub-communities 3-1, 3-2, and 3-3 and sub-communities 2-1, 2-2, 2-3, and 2-4, respectively. All of these sub-communities are also different and can be distinguished. In particular, although the regions corresponding to community numbers 2-1, 2-2, and 2-4 in Figure 6(b) are tiny, they are still divided into independent communities.

Similarly, according to Definitions 8 and 9, we overlaid the POI dataset with the spatial regions corresponding to the three-level hierarchical community structures, and the results of the low, medium, and high spatial resolution community partitions are shown in Figure 7(a)–(c), respectively. We can also interpret the cohesiveness of the spatial interactions within the community and the heterogeneity between communities through the visual analysis of these figures.

Figure 7 The overlay map of the three-level hierarchical community partitions with the POI dataset. (a) Low spatial resolution. (b) Medium spatial resolution. (c) High spatial resolution.
Figure 7

The overlay map of the three-level hierarchical community partitions with the POI dataset. (a) Low spatial resolution. (b) Medium spatial resolution. (c) High spatial resolution.

5.1.3 Quantitative interpretation of communities using the distribution of containing geographical features

For the communities in Figures 6(c) and 7(c), due to the large number of communities (32 communities), it is not easy to visualize the differences between the regions of the communities in terms of the spatial distribution patterns of the geographical features. Therefore, we further performed the experiments using statistics of the geographical features contained in the corresponding grids of the communities.

According to Definitions 7 and 11, we conducted a statistical analysis of the areas of the land use dataset contained by the corresponding grids of the communities and calculated the entropy value of each distribution. The results are shown in Figure 8. As can be seen from Figure 8(a), the statistical distribution of the land use dataset contained in the corresponding spatial regions of the three communities (Figure 7(a)) is entirely different. Specifically, the area percentage of the land use dataset for community 1 is unbalanced, and Arable land (annual crops) alone accounts for nearly 50%. Moreover, the area percentages of the land use dataset in communities 2 and 3 are relatively balanced. In addition, the most significant proportions of the land use dataset in communities 2 and 3 are also different: the former is Arable land (annual crops), while the latter is industrial, commercial, public, military, and private units.

Figure 8 Statistical distribution of land use dataset of the three-level hierarchical community partitions. (a) Statistical distribution patterns of the three communities in Figure 7(a). (b) Statistical distribution patterns of sub-communities in Figure 7(b). (c), (d), (e), (f), (g), and (h) are statistical distribution patterns of sub-communities 2–1, 2–3, 2–4, 3–1, 3–2, and 3–3, respectively
Figure 8

Statistical distribution of land use dataset of the three-level hierarchical community partitions. (a) Statistical distribution patterns of the three communities in Figure 7(a). (b) Statistical distribution patterns of sub-communities in Figure 7(b). (c), (d), (e), (f), (g), and (h) are statistical distribution patterns of sub-communities 2–1, 2–3, 2–4, 3–1, 3–2, and 3–3, respectively

In Figure 8(b), the statistical distribution patterns of sub-communities 2-1 and 2-2 and sub-communities 2-3 and 2-4 have differences. Similarly, we can also find the differences in the statistical distributions of the sub-communities 3-1, 3-2, and 3-3.

Due to the large number of communities included in the high spatial resolution community partitions, we grouped and compared their corresponding sub-communities in the medium spatial resolution community partitions. Specifically, the comparison of the statistical distribution of the land use dataset of sub-communities 2-1, 2-3, 2-4, 3-1, 3-2, and 3-3 in the medium spatial resolution community partitions is shown in Figure 8(c)–(h), respectively. All sub-community groups can be clearly distinguished by the statistical distribution of the land use dataset contained in the sub-communities’ corresponding grids.

Similarly, according to Definitions 10 and 12, we conducted statistical analysis on the POI dataset in terms of the point counts contained in the corresponding grids of the community and calculated the entropy value of each distribution. The statistical analysis results are shown in Figure 9, where Figure 9(a) and (b) are the statistical distribution patterns of communities and sub-communities in the low spatial resolution community partitions and the medium spatial resolution community partitions. To further distinguish the divisions of the high spatial resolution community partitions, the statistical distributions of the POI dataset were also performed group comparisons, which correspond to communities 2-1, 2-3, 2-4, 3-1, 3-2, and 3-3 in the medium spatial resolution community partitions. The results are shown in Figure 9(c)–(h), respectively.

Figure 9 Statistical distribution of the POI dataset of the three-level hierarchical community partitions. (a) Statistical distribution patterns of the three communities in Figure 7(a). (b) Statistical distribution patterns of sub-communities in Figure 7(b). (c), (d), (e), (f), (g), and (h) are statistical distribution patterns of sub-communities 2–1, 2–3, 2–4, 3–1, 3–2, and 3–3, respectively
Figure 9

Statistical distribution of the POI dataset of the three-level hierarchical community partitions. (a) Statistical distribution patterns of the three communities in Figure 7(a). (b) Statistical distribution patterns of sub-communities in Figure 7(b). (c), (d), (e), (f), (g), and (h) are statistical distribution patterns of sub-communities 2–1, 2–3, 2–4, 3–1, 3–2, and 3–3, respectively

Comparing Figures 8 and 9, we found that the opposite phenomena occur in the statistical distributions of the land use and POI dataset. In particular, for low spatial resolution community partitions in Figure 9(a) and the medium spatial resolution community partitions, the difference in the statistical distribution is not particularly obvious. In contrast, for the high spatial resolution community partitions, the difference is very distinct. That is, at low and medium spatial resolution scales, the statistical distributions of the POI dataset of the different communities are less heterogeneous. In contrast, at a high spatial resolution scale, the corresponding statistical distribution patterns are highly heterogeneous. This phenomenon is precisely the opposite of the results for the land use dataset.

This phenomenon is related to the time aggregation scale of the spatial interaction records and the characteristics of the land use dataset and the POI dataset. Specifically, in this experiment, we aggregated half a month’s worth of spatial interaction records to construct a spatial interaction network and detected communities from it. In this half a month, the POI dataset contained in the detected communities may change (i.e., the POI dataset is generally collected by some companies and is updated frequently), whereas the land use dataset contained is rarely updated (i.e., the land use dataset is usually released by the urban planning department of the government). Therefore, at low and medium spatial resolution scales, the distinctions between communities are more relevant to the statistical distribution of the land use dataset. In contrast, at a high spatial resolution scale, the differences between communities are more relevant to the statistical distribution of the POI dataset.

5.2 Quantitative interpretation of the stable and active regions found in the community time series

In this experiment, we aggregated each day’s spatial interaction records to construct 29 spatial interaction networks (data for November 24 is missing). Then, we utilized the Louvain method to detect sequences of communities in these 29 networks. As it is exhausting to identify the active and stable regions from the sequence of communities with high spatial resolution, we focused on the two sequences of communities with low and medium spatial resolution community partitions.

According to Definition 4, we discovered one active region and two stable regions in these two sequences of communities. The overlay maps of the regions with the sequences of communities in the low and medium spatial resolution community partitions are shown in Figures 10 and 11, respectively, where the active regions were marked by the purple circles and the stable regions were marked by the green circles.

Figure 10 The overlay maps of the regions with the sequence of communities in the low spatial resolution community partitions.
Figure 10

The overlay maps of the regions with the sequence of communities in the low spatial resolution community partitions.

Figure 11 The overlay maps of the regions with the sequence of communities in the medium spatial resolution community partitions.
Figure 11

The overlay maps of the regions with the sequence of communities in the medium spatial resolution community partitions.

According to Definitions 5, 6, 7, and 11, we conducted a statistical analysis of the area of the land use dataset contained in the one active region and two stable regions found in the lower and medium spatial resolution community time series, and the results are shown in Figures 12(a) and (b), respectively. As can be seen from these figures, there is a clear distinction between the statistical distributions of the land use dataset in the active region and the stable regions. That is, the three regions can be interpreted based on the distributions of the land use dataset. Specifically, the stable region has an unbalanced statistical distribution of land use. That is, the stable region is predominantly Arable land (annual crops), whereas the two active regions have more balanced land use distributions.

Figure 12 Statistical distribution of the land use dataset in the active regions and stable regions based on the sequence of communities. (a) Low spatial resolution. (b) Medium spatial resolution.
Figure 12

Statistical distribution of the land use dataset in the active regions and stable regions based on the sequence of communities. (a) Low spatial resolution. (b) Medium spatial resolution.

Similarly, according to Definitions 8, 9, 10, and 12, we conducted a statistical analysis of the counts of the POIs contained in the one active region and two stable regions found in the lower and medium spatial resolution community time series, and the results are shown in Figures 13(a) and (b), respectively. We found that obvious distinctions also exist in the statistical distributions of the POI dataset.

Figure 13 Statistical distribution of the POI dataset in the active regions and stable regions based on the sequence of communities. (a) Low spatial resolution. (b) Medium spatial resolution.
Figure 13

Statistical distribution of the POI dataset in the active regions and stable regions based on the sequence of communities. (a) Low spatial resolution. (b) Medium spatial resolution.

Finally, we concluded that for this dynamic method, both the land use dataset and the POI dataset could be utilized to clearly distinguish and quantitatively interpret the active regions and the stable regions. The active regions and the stable regions accurately capture the patterns of the spatial interactions reflected by the telephone calls.

6 Conclusions

As the existing methods for evaluating and interpreting the results of network community partitions are only qualitative descriptions, we proposed a quantitative and qualitative approach. This method quantifies and interprets the results of community partitions by overlaying the spatial regions corresponding to the detected communities with the related geographical features and calculating the distributions of the geographical features in the community regions. We selected Milan, Italy, as the study area, and the spatial interaction records reflected by the telephone calls, the land use dataset, and the POI dataset as the experimental data to perform extensive experiments to verify the proposed method. Experimental results demonstrated the effectiveness of our approach, which can quantitatively and qualitatively interpret network community partition results.

Acknowledgments

The authors would like to thank the anonymous referees and editor for their valuable comments, which significantly improved this article.

  1. Funding information: This research was supported by the Natural Science Foundation of China (grant number 41201465), the Postgraduate Research & Practice Innovation Program of Jiangsu Province (grant number KYCX21_0764 and KYCX21_0765)，and the Natural Science Foundation of Jiangsu Province (grant numbers BK2012439 and BE2016774).

  2. Author contributions: Conceptualization: Zhang Haitao. Data curation: Zhang Haitao, Ji Kang. Formal analysis: Zhang Haitao, Ji Kang. Funding acquisition: Zhang Haitao, Shen Huixian. Investigation: Zhang Haitao. Methodology: Zhang Haitao, Ji Kang. Project administration: Zhang Haitao. Resources: Zhang Haitao. Software: Shen Huixian, Song Rui. Supervision: Zhang Haitao. Validation: Ji Kang. Visualization: Zhang Haitao, Ji Kang. Writing–review & editing: Ji Kang, Liu Jin Yuan, Yang Yu Xin. All authors have read and agreed to the published version of the manuscript.

  3. Conflict of interest: All authors declare that there is no conflict of interest.

  4. Data availability statement: There are three types of datasets used in our research, including the Telephone dataset, land use dataset, and POI dataset. These data were derived from the following public domain resources: address: https://pan.baidu.com/s/1wF7Z_zjQcr2lCKwdqYn08w, Extraction code:3scu; https://land.copernicus.eu/local/urban-atlas/urban-atlas-2012?tab=download; https://www.openstreetmap.org/.

Appendix

Figure A1 An example of 13 spatial interaction records in a non-overlapping spatial tessellation with eight grids.
Figure A1

An example of 13 spatial interaction records in a non-overlapping spatial tessellation with eight grids.

Figure A2 The spatial interaction network corresponding to the spatial interaction records in Figure 1; for the convenience of calculation, we set the weight values to 1. In the experiment in Section 5, we use the intensity values as the weight values.
Figure A2

The spatial interaction network corresponding to the spatial interaction records in Figure 1; for the convenience of calculation, we set the weight values to 1. In the experiment in Section 5, we use the intensity values as the weight values.

Figure A3 Two communities were detected from the spatial interaction network in Figure 2, where the different colored nodes belong to different communities.
Figure A3

Two communities were detected from the spatial interaction network in Figure 2, where the different colored nodes belong to different communities.

Figure A4 (a) Grids corresponding to the communities in Figure 3, and (b) grids corresponding to the communities detected in the other different time interval spatial interaction network, where the colored grids belong to different communities. The grid 2 {{\rm{grid}}}_{2} , { grid 0 , grid 1 , grid 4 , and grid 5 } \{{{\rm{grid}}}_{0},{{\rm{grid}}}_{1},{{\rm{grid}}}_{4},{{\rm{and}}{\rm{grid}}}_{5}\} are, respectively, the active region and the stable regions corresponding to SINC = { v 0 , v 1 , v 2 , v 4 , v 5 } {\rm{SINC}}=\{{v}_{0},{v}_{1},{v}_{2},{v}_{4},{v}_{5}\} in Figure 4(a) and SINC = { v 0 , v 1 , v 4 , v 5 } {\rm{SINC}}=\{{v}_{0},{v}_{1},{v}_{4},{v}_{5}\} in Figure 4(b).
Figure A4

(a) Grids corresponding to the communities in Figure 3, and (b) grids corresponding to the communities detected in the other different time interval spatial interaction network, where the colored grids belong to different communities. The grid 2 , { grid 0 , grid 1 , grid 4 , and grid 5 } are, respectively, the active region and the stable regions corresponding to SINC = { v 0 , v 1 , v 2 , v 4 , v 5 } in Figure 4(a) and SINC = { v 0 , v 1 , v 4 , v 5 } in Figure 4(b).

Figure A5 An example of a land use dataset, which includes five land use types: arable land , urban fabric , water , forests , and airports {\rm{arable\; land}},{\rm{urban\; fabric}},{\rm{water}},{\rm{forests}},{\rm{and\; airports}} . That is, LandData = { ld 0 , ld 1 , ld 2 , ld 3 , ld 4 } {\rm{LandData}}=\{{{\rm{ld}}}_{0},{{\rm{ld}}}_{1}{,{\rm{ld}}}_{2},{{\rm{ld}}}_{3}{,{\rm{ld}}}_{4}\} , where ld 0 = ( geo 0 , arable land ) {{\rm{ld}}}_{0}=({{\rm{geo}}}_{0},{\rm{arable\; land}}) , ld 1 = ( { geo 1 , geo 3 } , urban fabric ) {{\rm{ld}}}_{1}=(\{{{\rm{geo}}}_{1},{{\rm{geo}}}_{3}\},{\rm{urban\; fabric}}) , ld 2 = ( { geo 2 , geo 4 , geo 5 } , water ) \hspace{1em}{{\rm{ld}}}_{2}=(\{{{\rm{geo}}}_{2},{{\rm{geo}}}_{4},{{\rm{geo}}}_{5}\},{\rm{water}}) , ld 3 = ( { geo 2 , geo 7 } , forests ) {{\rm{ld}}}_{3}=(\{{{\rm{geo}}}_{2},{{\rm{geo}}}_{7}\},{\rm{forests}}) , and ld 4 = ( { geo 3 , geo 7 } , airports ) {{\rm{ld}}}_{4}=(\{{{\rm{geo}}}_{3},{{\rm{geo}}}_{7}\},{\rm{airports}}) .
Figure A5

An example of a land use dataset, which includes five land use types: arable land , urban fabric , water , forests , and airports . That is, LandData = { ld 0 , ld 1 , ld 2 , ld 3 , ld 4 } , where ld 0 = ( geo 0 , arable land ) , ld 1 = ( { geo 1 , geo 3 } , urban fabric ) , ld 2 = ( { geo 2 , geo 4 , geo 5 } , water ) , ld 3 = ( { geo 2 , geo 7 } , forests ) , and ld 4 = ( { geo 3 , geo 7 } , airports ) .

Figure A6 Spatial grids correspond to the communities overlaid with the land use dataset in Figure 5, where SIN C Lan d OL 1 = { ol 0 , ol 1 , ol 2 , ol 4 , ol 5 } {{\rm{SIN}}{{\rm{C}}}_{{\rm{Lan}}{{\rm{d}}}_{{\rm{OL}}}}}_{1}=\{{{\rm{ol}}}_{0},{{\rm{ol}}}_{1}{,{\rm{ol}}}_{2}{,{\rm{ol}}}_{4},{{\rm{ol}}}_{5}\} , ol 0 = { v 0 , ( grid 0 , arable land ) } , ol 1 = { v 1 , ( grid 1 , urban fabric ) } , ol 2 = { v 2 , ( grid 2 , water , forests ) } , ol 4 = { v 4 , ( grid 4 , water , road ) } , and ol 5 = { v 5 , ( grid 5 , water ) } {{\rm{ol}}}_{0}=\{{v}_{0},({{\rm{grid}}}_{0},{\rm{arable\; land}})\},{{\rm{ol}}}_{1}=\{{v}_{1},({{\rm{grid}}}_{1},{\rm{urban\; fabric}})\},{{\rm{ol}}}_{2}=\hspace{0.33em}\{{v}_{2},({{\rm{grid}}}_{2},{\rm{water}},{\rm{forests}})\},{{\rm{ol}}}_{4}=\{{v}_{4},({{\rm{grid}}}_{4},{\rm{water}},{\rm{road}})\},{{\rm{and\; ol}}}_{5}=\{{v}_{5},({{\rm{grid}}}_{5},{\rm{water}})\} . SIN C Lan d OL 2 = { ol 3 , ol 6 , ol 7 } {{\rm{SIN}}{{\rm{C}}}_{{\rm{Lan}}{{\rm{d}}}_{{\rm{OL}}}}}_{2}=\{{{\rm{ol}}}_{3},{{\rm{ol}}}_{6}{,{\rm{ol}}}_{7}\} , where ol 3 = { v 3 , ( grid 3 , urban fabric , airports ) } {{\rm{ol}}}_{3}=\{{v}_{3},({{\rm{grid}}}_{3},{\rm{urban\; fabric}},{\rm{airports}})\} , ol 6 = { v 6 , ( grid 6 , road ) } {{\rm{ol}}}_{6}=\{{v}_{6},({{\rm{grid}}}_{6},{\rm{road}})\} , and ol 7 = { v 7 , ( grid 7 , forests , airports ) } . {{\rm{ol}}}_{7}=\{{v}_{7},({{\rm{grid}}}_{7},{\rm{forests}},{\rm{airports}})\}.
Figure A6

Spatial grids correspond to the communities overlaid with the land use dataset in Figure 5, where SIN C Lan d OL 1 = { ol 0 , ol 1 , ol 2 , ol 4 , ol 5 } , ol 0 = { v 0 , ( grid 0 , arable land ) } , ol 1 = { v 1 , ( grid 1 , urban fabric ) } , ol 2 = { v 2 , ( grid 2 , water , forests ) } , ol 4 = { v 4 , ( grid 4 , water , road ) } , and ol 5 = { v 5 , ( grid 5 , water ) } . SIN C Lan d OL 2 = { ol 3 , ol 6 , ol 7 } , where ol 3 = { v 3 , ( grid 3 , urban fabric , airports ) } , ol 6 = { v 6 , ( grid 6 , road ) } , and ol 7 = { v 7 , ( grid 7 , forests , airports ) } .

Figure A7 An example of a POI dataset, which includes seven POI types: shopping, recreation, sightseeing, catering,business,accommodation,and public service, i.e., POIData={pd0, pd1, pd2, pd3, pd4, pd5, pd6, pd7 }. Where pd0 = {(geo0, shopping), (geo0, recreation), (geo0, sightseeing)}, pd1 = {(geo1, recreation)}, pd2 = {(geo2, business), (geo2, accommodation)}, pd3 = {(geo3, recreation)}, pd4 = {(geo4, recreation)}, pd5 = {(geo5, public service), (geo5, business), (geo5, recreation)}, pd6 = {(geo6, sightseeing)}, and pd7 = {(geo7, public service), (geo7, sightseeing)}.
Figure A7

An example of a POI dataset, which includes seven POI types: shopping, recreation, sightseeing, catering,business,accommodation,and public service, i.e., POIData={pd0, pd1, pd2, pd3, pd4, pd5, pd6, pd7 }. Where pd0 = {(geo0, shopping), (geo0, recreation), (geo0, sightseeing)}, pd1 = {(geo1, recreation)}, pd2 = {(geo2, business), (geo2, accommodation)}, pd3 = {(geo3, recreation)}, pd4 = {(geo4, recreation)}, pd5 = {(geo5, public service), (geo5, business), (geo5, recreation)}, pd6 = {(geo6, sightseeing)}, and pd7 = {(geo7, public service), (geo7, sightseeing)}.

Figure A8 Spatial grids correspond to the communities overlaid with the POI dataset shown in Figure 7, where SINC_POI_OL 1 = { ol 0 , ol 1 , ol 2 , ol 4 , ol 5 } {{\rm{SINC\_POI\_OL}}}_{1}=\{{{\rm{ol}}}_{0},{{\rm{ol}}}_{1},{{\rm{ol}}}_{2}{,{\rm{ol}}}_{4}{,{\rm{ol}}}_{5}\} , ol 0 = { v 0 , ( grid 0 , ( shopping , recreation , sightseeing ) ) } , ol 1 = { v 1 , ( grid 1 , recreation ) } , ol 2 = { v 2 , ( grid 2 , ( business , accommodation ) ) } , ol 4 = { v 4 , ( grid 4 , recreation ) } , and ol 5 = { v 5 , ( grid 5 , ( business , public service , recreation ) ) } {{\rm{ol}}}_{0}=\{{v}_{0},({{\rm{grid}}}_{0},({\rm{shopping}},{\rm{recreation}},{\rm{sightseeing}}))\},{{\rm{ol}}}_{1}=\hspace{0.33em}\{{v}_{1},({{\rm{grid}}}_{1},{\rm{recreation}})\},{{\rm{ol}}}_{2}=\{{v}_{2},({{\rm{grid}}}_{2},({\rm{business}},{\rm{accommodation}}))\},{{\rm{ol}}}_{4}=\hspace{0.33em}\{{v}_{4},({{\rm{grid}}}_{4},{\rm{recreation}})\},{{\rm{and\; ol}}}_{5}=\{{v}_{5},({{\rm{grid}}}_{5},({\rm{business}},{\rm{public\; service}},{\rm{recreation}}))\} . SINC_POI_OL 2 = { ol 3 , ol 6 , ol 7 } {{\rm{SINC\_POI\_OL}}}_{2}=\{{{\rm{ol}}}_{3},{{\rm{ol}}}_{6}{,{\rm{ol}}}_{7}\} , ol 3 = { v 3 , ( grid 3 , recreation ) } , ol 6 = { v 6 , ( grid 6 , sightseeing ) } {{\rm{ol}}}_{3}=\{{v}_{3},({{\rm{grid}}}_{3},{\rm{recreation}})\},{{\rm{ol}}}_{6}=\{{v}_{6},({{\rm{grid}}}_{6},{\rm{sightseeing}})\} , and ol 7 = { v 7 , ( grid 7 , ( public service , catering ) ) } . {{ol}}_{7}=\{{v}_{7},({{\rm{grid}}}_{7},({\rm{public\; service}},{\rm{catering}}))\}.
Figure A8

Spatial grids correspond to the communities overlaid with the POI dataset shown in Figure 7, where SINC_POI_OL 1 = { ol 0 , ol 1 , ol 2 , ol 4 , ol 5 } , ol 0 = { v 0 , ( grid 0 , ( shopping , recreation , sightseeing ) ) } , ol 1 = { v 1 , ( grid 1 , recreation ) } , ol 2 = { v 2 , ( grid 2 , ( business , accommodation ) ) } , ol 4 = { v 4 , ( grid 4 , recreation ) } , and ol 5 = { v 5 , ( grid 5 , ( business , public service , recreation ) ) } . SINC_POI_OL 2 = { ol 3 , ol 6 , ol 7 } , ol 3 = { v 3 , ( grid 3 , recreation ) } , ol 6 = { v 6 , ( grid 6 , sightseeing ) } , and ol 7 = { v 7 , ( grid 7 , ( public service , catering ) ) } .

References

[1] Zhang H, Shen H, Ji K, Song R, Liu J, Yang Y. An urban hot/cold spot detection method based on the page rank value of spatial interaction networks constructed from human communication records. ISPRS Int J Geo-Inf. 2022;11(3):210.10.3390/ijgi11030210Suche in Google Scholar

[2] Sobolevsky S, Szell M, Campari R, Couronné T, Smoreda Z, Ratti C. Delineating geographical regions with networks of human interactions in an extensive set of countries. PLoS One. 2013;8(12):e81707.10.1371/journal.pone.0081707Suche in Google Scholar PubMed PubMed Central

[3] Blondel VD, Guillaume JL, Lambiotte R, Lefebvre E. Fast unfolding of communities in large networks. J Stat Mech Theory Exp. 2008;2008(10):P10008.10.1088/1742-5468/2008/10/P10008Suche in Google Scholar

[4] Newman M, Girvan M. Finding and evaluating community structure in networks. 2004;69(2):26113.10.1103/PhysRevE.69.026113Suche in Google Scholar PubMed

[5] Fu G, Hou C, Yao X. Learning topological representation for networks via hierarchical sampling; 2019. 10.1109/IJCNN.2019.8851893.Suche in Google Scholar

[6] Blondel V, Krings G, Thomas I. Regions and borders of mobile telephony in Belgium and in the Brussels Metropolitan zone. Bruss Stud. 2010;42(4):1–12.10.4000/brussels.806Suche in Google Scholar

[7] Ratti C, Sobolevsky S, Calabrese F, Andris C, Reades J. Redrawing the map of Great Britain from a network of human interactions. PLoS One. 2010;5(12):1–6.10.1371/journal.pone.0014248Suche in Google Scholar PubMed PubMed Central

[8] Kallus Z, N, Barankai, Szüle J, Vattay G. Spatial fingerprints of community structure in human interaction network for an extensive set of large-scale regions. PLoS One. 2015;10(5):e0126713.10.1371/journal.pone.0126713Suche in Google Scholar PubMed PubMed Central

[9] Gao S, Liu Y, Wang Y, Ma X. Discovering spatial interaction communities from mobile phone data. Trans GIS. 2013;17(3):463–81.10.1111/tgis.12042Suche in Google Scholar

[10] Botta F, Del Genio CI. Analysis of the communities of an urban mobile phone network. PLoS one. 2017;12(3):e0174198.10.1371/journal.pone.0174198Suche in Google Scholar PubMed PubMed Central

[11] Xu Y, Li J, Belyi A, Park S. Characterizing destination networks through mobility traces of international tourists—A case study using a nationwide mobile positioning dataset. Tour Manag. 2021;82:104195.10.1016/j.tourman.2020.104195Suche in Google Scholar

[12] Wang ZJ, Chen ZX, Wu JY, Yu HW, Yao XM. Detecting latent urban mobility structure using mobile phone data. Mod Phys Lett B. 2020;34(30):2050342.10.1142/S021798492050342XSuche in Google Scholar

[13] Chen Y, Xu J, Xu M. Finding community structure in spatially constrained complex networks. Int J Geogr Inf Sci. 2015;29(6):889–911.10.1080/13658816.2014.999244Suche in Google Scholar

[14] Koylu C. Discovering multi-scale community structures from the interpersonal communication network on Twitter; 201710.1007/978-3-319-65993-0_7Suche in Google Scholar

[15] Arenas A, Duch J, Fernández A, Gómez S. Size reduction of complex networks preserving modularity. N J Phys. 2007;176(9):1–14. 10.1088/1367-2630/9/6/176.Suche in Google Scholar

[16] Shannon CE. A mathematical theory of communication. Bell Syst Tech J. 1948;27(3):379–423.10.1002/j.1538-7305.1948.tb01338.xSuche in Google Scholar

[17] Spellerberg IF, Fedor PJ. A tribute to Claude Shannon (1916–2001) and a plea for more rigorous use of species richness, species diversity and the ‘Shannon–Wiener’ Index. Glob Ecol Biogeogr. 2003;12(3):177–9.10.1046/j.1466-822X.2003.00015.xSuche in Google Scholar

[18] Zhang H, Yu C, Jin Y. A novel method for classifying function of spatial regions based on two sets of characteristics indicated by trajectories. Int J Data Warehous Min. 2020;16(3):1–19.10.4018/IJDWM.2020070101Suche in Google Scholar
Received: 2022-03-24
Revised: 2023-02-01
Accepted: 2023-02-10
Published Online: 2023-11-07

© 2023 the author(s), published by De Gruyter

This work is licensed under the Creative Commons Attribution 4.0 International License.

Artikel in diesem Heft

  1. Regular Articles
  2. Diagenesis and evolution of deep tight reservoirs: A case study of the fourth member of Shahejie Formation (cg: 50.4-42 Ma) in Bozhong Sag
  3. Petrography and mineralogy of the Oligocene flysch in Ionian Zone, Albania: Implications for the evolution of sediment provenance and paleoenvironment
  4. Biostratigraphy of the Late Campanian–Maastrichtian of the Duwi Basin, Red Sea, Egypt
  5. Structural deformation and its implication for hydrocarbon accumulation in the Wuxia fault belt, northwestern Junggar basin, China
  6. Carbonate texture identification using multi-layer perceptron neural network
  7. Metallogenic model of the Hongqiling Cu–Ni sulfide intrusions, Central Asian Orogenic Belt: Insight from long-period magnetotellurics
  8. Assessments of recent Global Geopotential Models based on GPS/levelling and gravity data along coastal zones of Egypt
  9. Accuracy assessment and improvement of SRTM, ASTER, FABDEM, and MERIT DEMs by polynomial and optimization algorithm: A case study (Khuzestan Province, Iran)
  10. Uncertainty assessment of 3D geological models based on spatial diffusion and merging model
  11. Evaluation of dynamic behavior of varved clays from the Warsaw ice-dammed lake, Poland
  12. Impact of AMSU-A and MHS radiances assimilation on Typhoon Megi (2016) forecasting
  13. Contribution to the building of a weather information service for solar panel cleaning operations at Diass plant (Senegal, Western Sahel)
  14. Measuring spatiotemporal accessibility to healthcare with multimodal transport modes in the dynamic traffic environment
  15. Mathematical model for conversion of groundwater flow from confined to unconfined aquifers with power law processes
  16. NSP variation on SWAT with high-resolution data: A case study
  17. Reconstruction of paleoglacial equilibrium-line altitudes during the Last Glacial Maximum in the Diancang Massif, Northwest Yunnan Province, China
  18. A prediction model for Xiangyang Neolithic sites based on a random forest algorithm
  19. Determining the long-term impact area of coastal thermal discharge based on a harmonic model of sea surface temperature
  20. Origin of block accumulations based on the near-surface geophysics
  21. Investigating the limestone quarries as geoheritage sites: Case of Mardin ancient quarry
  22. Population genetics and pedigree geography of Trionychia japonica in the four mountains of Henan Province and the Taihang Mountains
  23. Performance audit evaluation of marine development projects based on SPA and BP neural network model
  24. Study on the Early Cretaceous fluvial-desert sedimentary paleogeography in the Northwest of Ordos Basin
  25. Detecting window line using an improved stacked hourglass network based on new real-world building façade dataset
  26. Automated identification and mapping of geological folds in cross sections
  27. Silicate and carbonate mixed shelf formation and its controlling factors, a case study from the Cambrian Canglangpu formation in Sichuan basin, China
  28. Ground penetrating radar and magnetic gradient distribution approach for subsurface investigation of solution pipes in post-glacial settings
  29. Research on pore structures of fine-grained carbonate reservoirs and their influence on waterflood development
  30. Risk assessment of rain-induced debris flow in the lower reaches of Yajiang River based on GIS and CF coupling models
  31. Multifractal analysis of temporal and spatial characteristics of earthquakes in Eurasian seismic belt
  32. Surface deformation and damage of 2022 (M 6.8) Luding earthquake in China and its tectonic implications
  33. Differential analysis of landscape patterns of land cover products in tropical marine climate zones – A case study in Malaysia
  34. DEM-based analysis of tectonic geomorphologic characteristics and tectonic activity intensity of the Dabanghe River Basin in South China Karst
  35. Distribution, pollution levels, and health risk assessment of heavy metals in groundwater in the main pepper production area of China
  36. Study on soil quality effect of reconstructing by Pisha sandstone and sand soil
  37. Understanding the characteristics of loess strata and quaternary climate changes in Luochuan, Shaanxi Province, China, through core analysis
  38. Dynamic variation of groundwater level and its influencing factors in typical oasis irrigated areas in Northwest China
  39. Creating digital maps for geotechnical characteristics of soil based on GIS technology and remote sensing
  40. Changes in the course of constant loading consolidation in soil with modeled granulometric composition contaminated with petroleum substances
  41. Correlation between the deformation of mineral crystal structures and fault activity: A case study of the Yingxiu-Beichuan fault and the Milin fault
  42. Cognitive characteristics of the Qiang religious culture and its influencing factors in Southwest China
  43. Spatiotemporal variation characteristics analysis of infrastructure iron stock in China based on nighttime light data
  44. Interpretation of aeromagnetic and remote sensing data of Auchi and Idah sheets of the Benin-arm Anambra basin: Implication of mineral resources
  45. Building element recognition with MTL-AINet considering view perspectives
  46. Characteristics of the present crustal deformation in the Tibetan Plateau and its relationship with strong earthquakes
  47. Influence of fractures in tight sandstone oil reservoir on hydrocarbon accumulation: A case study of Yanchang Formation in southeastern Ordos Basin
  48. Nutrient assessment and land reclamation in the Loess hills and Gulch region in the context of gully control
  49. Handling imbalanced data in supervised machine learning for lithological mapping using remote sensing and airborne geophysical data
  50. Spatial variation of soil nutrients and evaluation of cultivated land quality based on field scale
  51. Lignin analysis of sediments from around 2,000 to 1,000 years ago (Jiulong River estuary, southeast China)
  52. Assessing OpenStreetMap roads fitness-for-use for disaster risk assessment in developing countries: The case of Burundi
  53. Transforming text into knowledge graph: Extracting and structuring information from spatial development plans
  54. A symmetrical exponential model of soil temperature in temperate steppe regions of China
  55. A landslide susceptibility assessment method based on auto-encoder improved deep belief network
  56. Numerical simulation analysis of ecological monitoring of small reservoir dam based on maximum entropy algorithm
  57. Morphometry of the cold-climate Bory Stobrawskie Dune Field (SW Poland): Evidence for multi-phase Lateglacial aeolian activity within the European Sand Belt
  58. Adopting a new approach for finding missing people using GIS techniques: A case study in Saudi Arabia’s desert area
  59. Geological earthquake simulations generated by kinematic heterogeneous energy-based method: Self-arrested ruptures and asperity criterion
  60. Semi-automated classification of layered rock slopes using digital elevation model and geological map
  61. Geochemical characteristics of arc fractionated I-type granitoids of eastern Tak Batholith, Thailand
  62. Lithology classification of igneous rocks using C-band and L-band dual-polarization SAR data
  63. Analysis of artificial intelligence approaches to predict the wall deflection induced by deep excavation
  64. Evaluation of the current in situ stress in the middle Permian Maokou Formation in the Longnüsi area of the central Sichuan Basin, China
  65. Utilizing microresistivity image logs to recognize conglomeratic channel architectural elements of Baikouquan Formation in slope of Mahu Sag
  66. Resistivity cutoff of low-resistivity and low-contrast pays in sandstone reservoirs from conventional well logs: A case of Paleogene Enping Formation in A-Oilfield, Pearl River Mouth Basin, South China Sea
  67. Examining the evacuation routes of the sister village program by using the ant colony optimization algorithm
  68. Spatial objects classification using machine learning and spatial walk algorithm
  69. Study on the stabilization mechanism of aeolian sandy soil formation by adding a natural soft rock
  70. Bump feature detection of the road surface based on the Bi-LSTM
  71. The origin and evolution of the ore-forming fluids at the Manondo-Choma gold prospect, Kirk range, southern Malawi
  72. A retrieval model of surface geochemistry composition based on remotely sensed data
  73. Exploring the spatial dynamics of cultural facilities based on multi-source data: A case study of Nanjing’s art institutions
  74. Study of pore-throat structure characteristics and fluid mobility of Chang 7 tight sandstone reservoir in Jiyuan area, Ordos Basin
  75. Study of fracturing fluid re-discharge based on percolation experiments and sampling tests – An example of Fuling shale gas Jiangdong block, China
  76. Impacts of marine cloud brightening scheme on climatic extremes in the Tibetan Plateau
  77. Ecological protection on the West Coast of Taiwan Strait under economic zone construction: A case study of land use in Yueqing
  78. The time-dependent deformation and damage constitutive model of rock based on dynamic disturbance tests
  79. Evaluation of spatial form of rural ecological landscape and vulnerability of water ecological environment based on analytic hierarchy process
  80. Fingerprint of magma mixture in the leucogranites: Spectroscopic and petrochemical approach, Kalebalta-Central Anatolia, Türkiye
  81. Principles of self-calibration and visual effects for digital camera distortion
  82. UAV-based doline mapping in Brazilian karst: A cave heritage protection reconnaissance
  83. Evaluation and low carbon ecological urban–rural planning and construction based on energy planning mechanism
  84. Modified non-local means: A novel denoising approach to process gravity field data
  85. A novel travel route planning method based on an ant colony optimization algorithm
  86. Effect of time-variant NDVI on landside susceptibility: A case study in Quang Ngai province, Vietnam
  87. Regional tectonic uplift indicated by geomorphological parameters in the Bahe River Basin, central China
  88. Computer information technology-based green excavation of tunnels in complex strata and technical decision of deformation control
  89. Spatial evolution of coastal environmental enterprises: An exploration of driving factors in Jiangsu Province
  90. A comparative assessment and geospatial simulation of three hydrological models in urban basins
  91. Aquaculture industry under the blue transformation in Jiangsu, China: Structure evolution and spatial agglomeration
  92. Quantitative and qualitative interpretation of community partitions by map overlaying and calculating the distribution of related geographical features
  93. Numerical investigation of gravity-grouted soil-nail pullout capacity in sand
  94. Analysis of heavy pollution weather in Shenyang City and numerical simulation of main pollutants
  95. Road cut slope stability analysis for static and dynamic (pseudo-static analysis) loading conditions
  96. Forest biomass assessment combining field inventorying and remote sensing data
  97. Late Jurassic Haobugao granites from the southern Great Xing’an Range, NE China: Implications for postcollision extension of the Mongol–Okhotsk Ocean
  98. Petrogenesis of the Sukadana Basalt based on petrology and whole rock geochemistry, Lampung, Indonesia: Geodynamic significances
  99. Numerical study on the group wall effect of nodular diaphragm wall foundation in high-rise buildings
  100. Water resources utilization and tourism environment assessment based on water footprint
  101. Geochemical evaluation of the carbonaceous shale associated with the Permian Mikambeni Formation of the Tuli Basin for potential gas generation, South Africa
  102. Detection and characterization of lineaments using gravity data in the south-west Cameroon zone: Hydrogeological implications
  103. Study on spatial pattern of tourism landscape resources in county cities of Yangtze River Economic Belt
  104. The effect of weathering on drillability of dolomites
  105. Noise masking of near-surface scattering (heterogeneities) on subsurface seismic reflectivity
  106. Query optimization-oriented lateral expansion method of distributed geological borehole database
  107. Petrogenesis of the Morobe Granodiorite and their shoshonitic mafic microgranular enclaves in Maramuni arc, Papua New Guinea
  108. Environmental health risk assessment of urban water sources based on fuzzy set theory
  109. Spatial distribution of urban basic education resources in Shanghai: Accessibility and supply-demand matching evaluation
  110. Spatiotemporal changes in land use and residential satisfaction in the Huai River-Gaoyou Lake Rim area
  111. Walkaway vertical seismic profiling first-arrival traveltime tomography with velocity structure constraints
  112. Study on the evaluation system and risk factor traceability of receiving water body
  113. Predicting copper-polymetallic deposits in Kalatag using the weight of evidence model and novel data sources
  114. Temporal dynamics of green urban areas in Romania. A comparison between spatial and statistical data
  115. Passenger flow forecast of tourist attraction based on MACBL in LBS big data environment
  116. Varying particle size selectivity of soil erosion along a cultivated catena
  117. Relationship between annual soil erosion and surface runoff in Wadi Hanifa sub-basins
  118. Influence of nappe structure on the Carboniferous volcanic reservoir in the middle of the Hongche Fault Zone, Junggar Basin, China
  119. Dynamic analysis of MSE wall subjected to surface vibration loading
  120. Pre-collisional architecture of the European distal margin: Inferences from the high-pressure continental units of central Corsica (France)
  121. The interrelation of natural diversity with tourism in Kosovo
  122. Assessment of geosites as a basis for geotourism development: A case study of the Toplica District, Serbia
  123. IG-YOLOv5-based underwater biological recognition and detection for marine protection
  124. Monitoring drought dynamics using remote sensing-based combined drought index in Ergene Basin, Türkiye
  125. Review Articles
  126. The actual state of the geodetic and cartographic resources and legislation in Poland
  127. Evaluation studies of the new mining projects
  128. Comparison and significance of grain size parameters of the Menyuan loess calculated using different methods
  129. Scientometric analysis of flood forecasting for Asia region and discussion on machine learning methods
  130. Rainfall-induced transportation embankment failure: A review
  131. Rapid Communication
  132. Branch fault discovered in Tangshan fault zone on the Kaiping-Guye boundary, North China
  133. Technical Note
  134. Introducing an intelligent multi-level retrieval method for mineral resource potential evaluation result data
  135. Erratum
  136. Erratum to “Forest cover assessment using remote-sensing techniques in Crete Island, Greece”
  137. Addendum
  138. The relationship between heat flow and seismicity in global tectonically active zones
  139. Commentary
  140. Improved entropy weight methods and their comparisons in evaluating the high-quality development of Qinghai, China
  141. Special Issue: Geoethics 2022 - Part II
  142. Loess and geotourism potential of the Braničevo District (NE Serbia): From overexploitation to paleoclimate interpretation
https://doi.org/10.1515/geo-2022-0462
Schlagwörter für diesen Artikel
spatial interaction network; community partitions; geographical features; statistical distribution
Creative Commons
BY 4.0

Artikel in diesem Heft

  1. Regular Articles
  2. Diagenesis and evolution of deep tight reservoirs: A case study of the fourth member of Shahejie Formation (cg: 50.4-42 Ma) in Bozhong Sag
  3. Petrography and mineralogy of the Oligocene flysch in Ionian Zone, Albania: Implications for the evolution of sediment provenance and paleoenvironment
  4. Biostratigraphy of the Late Campanian–Maastrichtian of the Duwi Basin, Red Sea, Egypt
  5. Structural deformation and its implication for hydrocarbon accumulation in the Wuxia fault belt, northwestern Junggar basin, China
  6. Carbonate texture identification using multi-layer perceptron neural network
  7. Metallogenic model of the Hongqiling Cu–Ni sulfide intrusions, Central Asian Orogenic Belt: Insight from long-period magnetotellurics
  8. Assessments of recent Global Geopotential Models based on GPS/levelling and gravity data along coastal zones of Egypt
  9. Accuracy assessment and improvement of SRTM, ASTER, FABDEM, and MERIT DEMs by polynomial and optimization algorithm: A case study (Khuzestan Province, Iran)
  10. Uncertainty assessment of 3D geological models based on spatial diffusion and merging model
  11. Evaluation of dynamic behavior of varved clays from the Warsaw ice-dammed lake, Poland
  12. Impact of AMSU-A and MHS radiances assimilation on Typhoon Megi (2016) forecasting
  13. Contribution to the building of a weather information service for solar panel cleaning operations at Diass plant (Senegal, Western Sahel)
  14. Measuring spatiotemporal accessibility to healthcare with multimodal transport modes in the dynamic traffic environment
  15. Mathematical model for conversion of groundwater flow from confined to unconfined aquifers with power law processes
  16. NSP variation on SWAT with high-resolution data: A case study
  17. Reconstruction of paleoglacial equilibrium-line altitudes during the Last Glacial Maximum in the Diancang Massif, Northwest Yunnan Province, China
  18. A prediction model for Xiangyang Neolithic sites based on a random forest algorithm
  19. Determining the long-term impact area of coastal thermal discharge based on a harmonic model of sea surface temperature
  20. Origin of block accumulations based on the near-surface geophysics
  21. Investigating the limestone quarries as geoheritage sites: Case of Mardin ancient quarry
  22. Population genetics and pedigree geography of Trionychia japonica in the four mountains of Henan Province and the Taihang Mountains
  23. Performance audit evaluation of marine development projects based on SPA and BP neural network model
  24. Study on the Early Cretaceous fluvial-desert sedimentary paleogeography in the Northwest of Ordos Basin
  25. Detecting window line using an improved stacked hourglass network based on new real-world building façade dataset
  26. Automated identification and mapping of geological folds in cross sections
  27. Silicate and carbonate mixed shelf formation and its controlling factors, a case study from the Cambrian Canglangpu formation in Sichuan basin, China
  28. Ground penetrating radar and magnetic gradient distribution approach for subsurface investigation of solution pipes in post-glacial settings
  29. Research on pore structures of fine-grained carbonate reservoirs and their influence on waterflood development
  30. Risk assessment of rain-induced debris flow in the lower reaches of Yajiang River based on GIS and CF coupling models
  31. Multifractal analysis of temporal and spatial characteristics of earthquakes in Eurasian seismic belt
  32. Surface deformation and damage of 2022 (M 6.8) Luding earthquake in China and its tectonic implications
  33. Differential analysis of landscape patterns of land cover products in tropical marine climate zones – A case study in Malaysia
  34. DEM-based analysis of tectonic geomorphologic characteristics and tectonic activity intensity of the Dabanghe River Basin in South China Karst
  35. Distribution, pollution levels, and health risk assessment of heavy metals in groundwater in the main pepper production area of China
  36. Study on soil quality effect of reconstructing by Pisha sandstone and sand soil
  37. Understanding the characteristics of loess strata and quaternary climate changes in Luochuan, Shaanxi Province, China, through core analysis
  38. Dynamic variation of groundwater level and its influencing factors in typical oasis irrigated areas in Northwest China
  39. Creating digital maps for geotechnical characteristics of soil based on GIS technology and remote sensing
  40. Changes in the course of constant loading consolidation in soil with modeled granulometric composition contaminated with petroleum substances
  41. Correlation between the deformation of mineral crystal structures and fault activity: A case study of the Yingxiu-Beichuan fault and the Milin fault
  42. Cognitive characteristics of the Qiang religious culture and its influencing factors in Southwest China
  43. Spatiotemporal variation characteristics analysis of infrastructure iron stock in China based on nighttime light data
  44. Interpretation of aeromagnetic and remote sensing data of Auchi and Idah sheets of the Benin-arm Anambra basin: Implication of mineral resources
  45. Building element recognition with MTL-AINet considering view perspectives
  46. Characteristics of the present crustal deformation in the Tibetan Plateau and its relationship with strong earthquakes
  47. Influence of fractures in tight sandstone oil reservoir on hydrocarbon accumulation: A case study of Yanchang Formation in southeastern Ordos Basin
  48. Nutrient assessment and land reclamation in the Loess hills and Gulch region in the context of gully control
  49. Handling imbalanced data in supervised machine learning for lithological mapping using remote sensing and airborne geophysical data
  50. Spatial variation of soil nutrients and evaluation of cultivated land quality based on field scale
  51. Lignin analysis of sediments from around 2,000 to 1,000 years ago (Jiulong River estuary, southeast China)
  52. Assessing OpenStreetMap roads fitness-for-use for disaster risk assessment in developing countries: The case of Burundi
  53. Transforming text into knowledge graph: Extracting and structuring information from spatial development plans
  54. A symmetrical exponential model of soil temperature in temperate steppe regions of China
  55. A landslide susceptibility assessment method based on auto-encoder improved deep belief network
  56. Numerical simulation analysis of ecological monitoring of small reservoir dam based on maximum entropy algorithm
  57. Morphometry of the cold-climate Bory Stobrawskie Dune Field (SW Poland): Evidence for multi-phase Lateglacial aeolian activity within the European Sand Belt
  58. Adopting a new approach for finding missing people using GIS techniques: A case study in Saudi Arabia’s desert area
  59. Geological earthquake simulations generated by kinematic heterogeneous energy-based method: Self-arrested ruptures and asperity criterion
  60. Semi-automated classification of layered rock slopes using digital elevation model and geological map
  61. Geochemical characteristics of arc fractionated I-type granitoids of eastern Tak Batholith, Thailand
  62. Lithology classification of igneous rocks using C-band and L-band dual-polarization SAR data
  63. Analysis of artificial intelligence approaches to predict the wall deflection induced by deep excavation
  64. Evaluation of the current in situ stress in the middle Permian Maokou Formation in the Longnüsi area of the central Sichuan Basin, China
  65. Utilizing microresistivity image logs to recognize conglomeratic channel architectural elements of Baikouquan Formation in slope of Mahu Sag
  66. Resistivity cutoff of low-resistivity and low-contrast pays in sandstone reservoirs from conventional well logs: A case of Paleogene Enping Formation in A-Oilfield, Pearl River Mouth Basin, South China Sea
  67. Examining the evacuation routes of the sister village program by using the ant colony optimization algorithm
  68. Spatial objects classification using machine learning and spatial walk algorithm
  69. Study on the stabilization mechanism of aeolian sandy soil formation by adding a natural soft rock
  70. Bump feature detection of the road surface based on the Bi-LSTM
  71. The origin and evolution of the ore-forming fluids at the Manondo-Choma gold prospect, Kirk range, southern Malawi
  72. A retrieval model of surface geochemistry composition based on remotely sensed data
  73. Exploring the spatial dynamics of cultural facilities based on multi-source data: A case study of Nanjing’s art institutions
  74. Study of pore-throat structure characteristics and fluid mobility of Chang 7 tight sandstone reservoir in Jiyuan area, Ordos Basin
  75. Study of fracturing fluid re-discharge based on percolation experiments and sampling tests – An example of Fuling shale gas Jiangdong block, China
  76. Impacts of marine cloud brightening scheme on climatic extremes in the Tibetan Plateau
  77. Ecological protection on the West Coast of Taiwan Strait under economic zone construction: A case study of land use in Yueqing
  78. The time-dependent deformation and damage constitutive model of rock based on dynamic disturbance tests
  79. Evaluation of spatial form of rural ecological landscape and vulnerability of water ecological environment based on analytic hierarchy process
  80. Fingerprint of magma mixture in the leucogranites: Spectroscopic and petrochemical approach, Kalebalta-Central Anatolia, Türkiye
  81. Principles of self-calibration and visual effects for digital camera distortion
  82. UAV-based doline mapping in Brazilian karst: A cave heritage protection reconnaissance
  83. Evaluation and low carbon ecological urban–rural planning and construction based on energy planning mechanism
  84. Modified non-local means: A novel denoising approach to process gravity field data
  85. A novel travel route planning method based on an ant colony optimization algorithm
  86. Effect of time-variant NDVI on landside susceptibility: A case study in Quang Ngai province, Vietnam
  87. Regional tectonic uplift indicated by geomorphological parameters in the Bahe River Basin, central China
  88. Computer information technology-based green excavation of tunnels in complex strata and technical decision of deformation control
  89. Spatial evolution of coastal environmental enterprises: An exploration of driving factors in Jiangsu Province
  90. A comparative assessment and geospatial simulation of three hydrological models in urban basins
  91. Aquaculture industry under the blue transformation in Jiangsu, China: Structure evolution and spatial agglomeration
  92. Quantitative and qualitative interpretation of community partitions by map overlaying and calculating the distribution of related geographical features
  93. Numerical investigation of gravity-grouted soil-nail pullout capacity in sand
  94. Analysis of heavy pollution weather in Shenyang City and numerical simulation of main pollutants
  95. Road cut slope stability analysis for static and dynamic (pseudo-static analysis) loading conditions
  96. Forest biomass assessment combining field inventorying and remote sensing data
  97. Late Jurassic Haobugao granites from the southern Great Xing’an Range, NE China: Implications for postcollision extension of the Mongol–Okhotsk Ocean
  98. Petrogenesis of the Sukadana Basalt based on petrology and whole rock geochemistry, Lampung, Indonesia: Geodynamic significances
  99. Numerical study on the group wall effect of nodular diaphragm wall foundation in high-rise buildings
  100. Water resources utilization and tourism environment assessment based on water footprint
  101. Geochemical evaluation of the carbonaceous shale associated with the Permian Mikambeni Formation of the Tuli Basin for potential gas generation, South Africa
  102. Detection and characterization of lineaments using gravity data in the south-west Cameroon zone: Hydrogeological implications
  103. Study on spatial pattern of tourism landscape resources in county cities of Yangtze River Economic Belt
  104. The effect of weathering on drillability of dolomites
  105. Noise masking of near-surface scattering (heterogeneities) on subsurface seismic reflectivity
  106. Query optimization-oriented lateral expansion method of distributed geological borehole database
  107. Petrogenesis of the Morobe Granodiorite and their shoshonitic mafic microgranular enclaves in Maramuni arc, Papua New Guinea
  108. Environmental health risk assessment of urban water sources based on fuzzy set theory
  109. Spatial distribution of urban basic education resources in Shanghai: Accessibility and supply-demand matching evaluation
  110. Spatiotemporal changes in land use and residential satisfaction in the Huai River-Gaoyou Lake Rim area
  111. Walkaway vertical seismic profiling first-arrival traveltime tomography with velocity structure constraints
  112. Study on the evaluation system and risk factor traceability of receiving water body
  113. Predicting copper-polymetallic deposits in Kalatag using the weight of evidence model and novel data sources
  114. Temporal dynamics of green urban areas in Romania. A comparison between spatial and statistical data
  115. Passenger flow forecast of tourist attraction based on MACBL in LBS big data environment
  116. Varying particle size selectivity of soil erosion along a cultivated catena
  117. Relationship between annual soil erosion and surface runoff in Wadi Hanifa sub-basins
  118. Influence of nappe structure on the Carboniferous volcanic reservoir in the middle of the Hongche Fault Zone, Junggar Basin, China
  119. Dynamic analysis of MSE wall subjected to surface vibration loading
  120. Pre-collisional architecture of the European distal margin: Inferences from the high-pressure continental units of central Corsica (France)
  121. The interrelation of natural diversity with tourism in Kosovo
  122. Assessment of geosites as a basis for geotourism development: A case study of the Toplica District, Serbia
  123. IG-YOLOv5-based underwater biological recognition and detection for marine protection
  124. Monitoring drought dynamics using remote sensing-based combined drought index in Ergene Basin, Türkiye
  125. Review Articles
  126. The actual state of the geodetic and cartographic resources and legislation in Poland
  127. Evaluation studies of the new mining projects
  128. Comparison and significance of grain size parameters of the Menyuan loess calculated using different methods
  129. Scientometric analysis of flood forecasting for Asia region and discussion on machine learning methods
  130. Rainfall-induced transportation embankment failure: A review
  131. Rapid Communication
  132. Branch fault discovered in Tangshan fault zone on the Kaiping-Guye boundary, North China
  133. Technical Note
  134. Introducing an intelligent multi-level retrieval method for mineral resource potential evaluation result data
  135. Erratum
  136. Erratum to “Forest cover assessment using remote-sensing techniques in Crete Island, Greece”
  137. Addendum
  138. The relationship between heat flow and seismicity in global tectonically active zones
  139. Commentary
  140. Improved entropy weight methods and their comparisons in evaluating the high-quality development of Qinghai, China
  141. Special Issue: Geoethics 2022 - Part II
  142. Loess and geotourism potential of the Braničevo District (NE Serbia): From overexploitation to paleoclimate interpretation
Jetzt anmelden und 20% sparen
Institutioneller Zugang
Wie funktioniert Authentifizierung?
Haben Sie eine Idee, wie wir unsere Website verbessern können?
Bitte schreiben Sie uns.
© 2025 De Gruyter Brill
Heruntergeladen am 8.9.2025 von https://www.degruyterbrill.com/document/doi/10.1515/geo-2022-0462/html
Button zum nach oben scrollen