Measuring Spatiotemporal Civil War Dimensions Using Community-Based Dynamic Network Representation (CoDNet)

How to use it and examples

This is a notebook showing the code of the proposed method for CoDNet, which was employed for the analysis of Spatiotemporal Civil [1]. The code is in Python 3.9, and some toolboxes are necessary to run the commands.

[1] Vega-Oliveros, D. and Koren, O., "Measuring Spatiotemporal Civil War Dimensions Using Community-Based Dynamic Network Representation (CoDNet)," in IEEE Transactions on Computational Social Systems, vol. 11, no. 1, pp. 1506-1516, Feb. 2024, doi: 10.1109/TCSS.2023.3241173.

[2] Koren, Ore, 2023, "Measuring Spatio-Temporal Civil War Dimensions Using Community-Based Dynamic Network Representation (CoDNet)", https://doi.org/10.7910/DVN/0S9AFT, Harvard Dataverse, V1

The following very common packages are necessary for running the code:

• numpy
• pandas
• networkx
• sklearn
• basemap
• igraph
• leidenalg
• networkx

We provide the next environment installation commands (in Anaconda) to facilitate reproducibility:

conda create -n codnet python=3.9
conda activate codnet
conda install conda-forge::basemap
conda install anaconda::scikit-learn
conda install -c conda-forge python-igraph
conda install conda-forge::leidenalg
conda install anaconda::networkx

After cloning the repository, the first step is to extract the files from the compressed folder data.7z. Then, make sure that you have in the root repository the data folder, which contains three files:

• conf_data.csv
• gridtemp8914.csv
• prec_dat.csv

import matplotlib.pyplot as plt
import sys,os
import numpy as np
import pandas as pd
import networkx as nx
import graphConstruction as gc
from codnet import codnet
from communitiesTS import communityTS

event = 'state'    
neighbors = 3
timeWindow = 'month' #For month or year sliding window
MAX_DISTANCE = 500 #MAX_DISTANCE=[100, 250, 500] km
#cols columns of metrics: ['n',"n_norm_comm","m","<k>","<s>","lo",'tmp_i','pre_i','tmp_mean_community','pre_mean_community'] 
cols =  ["<k>",'n', "m"] 
codnets = codnet()

gc.contructYearsNetworks(event,neighbors,MAX_DISTANCE,timeWindow = timeWindow)
codnets.contructConsolidateYearNetwork(neighbors, MAX_DISTANCE, timeWindow = timeWindow)

Creating 8 workers processing  4  jobs
net-year_1989_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_1990_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_1991_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_1993_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_1992_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_1994_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2001_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_1997_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_1995_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_1998_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2002_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_1996_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2003_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_1999_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2005_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2009_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2000_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2006_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2004_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2010_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2013_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2017_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2007_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2011_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2014_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2008_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2012_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2015_neighbors_3_MAX_DISTANCE_500_month_.gml
net-year_2016_neighbors_3_MAX_DISTANCE_500_month_.gml

codnets.printYearsNetworksfromGmlFiles(neighbors=3,timeWindow=timeWindow,MAX_DISTANCE=MAX_DISTANCE)
codnets.printYearsNetworksfromGmlFiles(neighbors=3,timeWindow=timeWindow,MAX_DISTANCE=MAX_DISTANCE,Community=False)

Creating 8 workers processing  4  jobs
G:\Meu Drive\DaVo\COLABORA\ORE\Code3\public\net-All-Years\net-ALL-years_neighbors_3_MAX_DISTANCE_500_month_.gml
MaxDegree  85
Creating 8 workers processing  4  jobs

commTS = communityTS()  #This is the codnet table of graph community time-series analyses
codnet_table = commTS.read_TableTSCells(timeWindow,MAX_DISTANCE,neighbors)

display(codnet_table)

<style scoped> .dataframe tbody tr th:only-of-type { vertical-align: middle; } </style>

	gid	cLabel	n	n_norm_comm	m	<k>	<s>	lo	tmp_i	pre_i	tmp_mean_community	pre_mean_community	year
0	138446	0	24.0	0.100000	38.0	3.166667	5.583333	0.246537	27.44917	3881.5910	24.525613	2450.489188	1989
1	137007	0	24.0	0.100000	38.0	3.166667	5.583333	0.246537	25.95000	3881.5910	24.525613	2450.489188	1989
2	144935	0	24.0	0.100000	38.0	3.166667	5.583333	0.246537	25.91416	1292.0700	24.525613	2450.489188	1989
3	144934	0	24.0	0.100000	38.0	3.166667	5.583333	0.246537	24.61666	1292.0700	24.525613	2450.489188	1989
4	137738	0	24.0	0.100000	38.0	3.166667	5.583333	0.246537	26.13500	1942.5240	24.525613	2450.489188	1989
...	...	...	...	...	...	...	...	...	...	...	...	...	...
94349	172576	9	61.0	0.462121	189.0	6.196721	25.868852	0.272221	24.38417	396.0897	18.560438	357.534235	2011
94350	166805	9	61.0	0.462121	189.0	6.196721	25.868852	0.272221	26.64001	154.0747	18.560438	357.534235	2011
94351	172575	9	61.0	0.462121	189.0	6.196721	25.868852	0.272221	18.28666	317.6752	18.560438	357.534235	2011
94352	171858	9	61.0	0.462121	189.0	6.196721	25.868852	0.272221	20.78417	396.0897	18.560438	357.534235	2011
94353	173305	9	61.0	0.462121	189.0	6.196721	25.868852	0.272221	27.18917	488.7719	18.560438	357.534235	2011

94354 rows × 13 columns

community_selection = [4,10,13,17,18,23]
fileGML = gc.getFileName(begin='net-ALL-years',neighbors=neighbors,
                               MAX_DISTANCE=MAX_DISTANCE, 
                               timeWindow=timeWindow, ext='gml',path=os.path.join(os.curdir, 'net-All-Years'))
 
G = nx.read_gml(fileGML) 
codnets.drawAllTimeCommunityGraph(Gy=G, figname="temp.png", min_comm_size=10, selection= community_selection)
commTS.draw_TSMapComunity(cols, timeWindow, MAX_DISTANCE, neighbors, community_selection)

​
​

corrDf = commTS.drawTs(cols, timeWindow,MAX_DISTANCE,neighbors)

​
​

#"Pearson (T)":rT, 
#'Pearson p-value (T)':rpvT,
#'Spearman (T)':sT,
#'Spearman p-value (T)':spvT,
#'Pearson (P)':rP,
#'Pearson p-value (P)':rpvP,
#'Spearman (P)':sP,
#'Spearman p-value (P)':spvP
display(corrDf)                              

<style scoped> .dataframe tbody tr th:only-of-type { vertical-align: middle; } </style>

	cLabel	size	timeWindow	Distance	neighbors	measure	rT	rpvT	sT	spvT	rP	rpvP	sP	spvP
0	0	25	month	500	3	<k>	-0.106347	0.612891	-0.057703	0.784104	0.013275	0.949784	0.094249	0.654068
1	0	25	month	500	3	n	-0.025214	0.904773	-0.030811	0.883763	-0.037229	0.859761	0.073946	0.725377
2	0	25	month	500	3	m	0.026501	0.899936	-0.075784	0.718819	-0.038396	0.855410	0.080400	0.702432
3	1	25	month	500	3	<k>	-0.140377	0.503318	-0.168462	0.420842	0.137318	0.512756	-0.007692	0.970889
4	1	25	month	500	3	n	-0.298122	0.147770	-0.255868	0.217010	0.288571	0.161827	0.244325	0.239186
...	...	...	...	...	...	...	...	...	...	...	...	...	...	...
79	32	17	month	500	3	n	-0.275030	0.285350	-0.295744	0.249109	-0.163956	0.529483	-0.088598	0.735256
80	32	17	month	500	3	m	-0.226075	0.382932	-0.286250	0.265343	-0.166783	0.522308	-0.111045	0.671352
81	33	11	month	500	3	<k>	0.004894	0.988607	-0.188359	0.579136	-0.186579	0.582793	-0.282539	0.399891
82	33	11	month	500	3	n	-0.228749	0.498684	-0.419718	0.198750	-0.228845	0.498499	-0.317121	0.342000
83	33	11	month	500	3	m	-0.142654	0.675637	-0.390803	0.234678	-0.220663	0.514387	-0.283798	0.397702

84 rows × 14 columns

# corr_measure: "r" for Parson and "s" for Spearman
# climate_measure: 'T' for temperature and 'P' for precipitation
commTS.draw_TSMap(corrDf, corr_measure="r", climate_measure='T')
commTS.draw_TSMap(corrDf, corr_measure="s", climate_measure='T')
commTS.draw_TSMap(corrDf, corr_measure="r", climate_measure='P')
commTS.draw_TSMap(corrDf, corr_measure="s", climate_measure='P')

Pearson correlation with all-period Codnet constructed with time window month, max.distance 500, neighbors 3

Significant communities (p-value < 0.05) with T:

 Communties: [4, 10, 13, 18] for metric <k>
 Communties: [4, 11, 13, 18, 21] for metric n
 Communties: [4, 13, 18] for metric m

Spearman correlation with all-period Codnet constructed with time window month, max.distance 500, neighbors 3

Significant communities (p-value < 0.05) with T:

 Communties: [4, 10, 13, 18] for metric <k>
 Communties: [4, 10, 11, 13, 18, 21, 26] for metric n
 Communties: [4, 10, 13, 18, 21] for metric m

​
Pearson correlation with all-period Codnet constructed with time window month, max.distance 500, neighbors 3

Significant communities (p-value < 0.05) with P:

 Communties: [17] for metric <k>
 Communties: [17, 23] for metric n
 Communties: [11, 23] for metric m

​
Spearman correlation with all-period Codnet constructed with time window month, max.distance 500, neighbors 3

Significant communities (p-value < 0.05) with P:

 Communties: [17, 23] for metric <k>
 Communties: [17, 23] for metric n
 Communties: [17, 23] for metric m

​
​

References

You can use this code as it is for academic purpose. If you found it useful for your research, we appreciate your reference to our work Measuring Spatiotemporal Civil War Dimensions Using Community-Based Dynamic Network Representation (CoDNet).

More information, supplemental material, original dataset, and the ungated version of the article, you can find it in https://www.orekoren.com/publications

[1] Vega-Oliveros, D. and Koren, O., "Measuring Spatiotemporal Civil War Dimensions Using Community-Based Dynamic Network Representation (CoDNet)" in IEEE Transactions on Computational Social Systems, vol. 11, no. 1, pp. 1506-1516, Feb. 2024. https://doi.org/10.1109/TCSS.2023.3241173.