{"id":141,"date":"2018-07-09T21:39:51","date_gmt":"2018-07-09T21:39:51","guid":{"rendered":"http:\/\/sarahjpurcell.sites.grinnell.edu\/sandbox_clone\/?page_id=141"},"modified":"2018-11-04T20:12:17","modified_gmt":"2018-11-04T20:12:17","slug":"network-analysis-ii","status":"publish","type":"page","link":"https:\/\/sarahjpurcell.sites.grinnell.edu\/digital_methods\/tutorials\/network-analysis-ii\/","title":{"rendered":"Network Analysis-Part III (NetworkX)"},"content":{"rendered":"

This tutorial was written by Katherine Walden<\/a>, Digital Liberal Arts Specialist at Grinnell College.<\/p>\n

This tutorial was reviewed by Sarah Purcell<\/a> (L.F. Parker Professor of History) and Gina Donovan<\/a> (Instructional Technologist) at Grinnell College, and edited by Papa Ampim-Darko, a student research assistant at Grinnell College.<\/p>\n

This tutorial is adapted from the Programming Historian\u2019s Exploring and Analyzing Network Data with Python tutorial<\/a>.<\/p>\n

\"Creative<\/a>
\nNetwork Analysis-Part II (NetworkX) is licensed under a Creative Commons Attribution-NonCommercial 4.0 International License<\/a>.<\/p>\n

\n

GUI interfaces like Palladio and Gephi help researchers use digital tools quickly without requiring extensive technical documentation. But, larger data sets or more complex analysis often goes beyond the capacity of a GUI tool. In the next section of the tutorial, we will be using NetworkX<\/a>, a Python package, to analyze and study more complex aspects of the Quaker data. NetworkX is a free Python library developed and released in 2005.<\/em><\/p>\n

\n
Getting the Data<\/h5>\n

We will be working with sample networked data sets based on the Oxford Dictionary of National Biography<\/a> and the Six Degrees of Francis Bacon<\/a> project. These data sets include a list of names and relationships for early seventeenth-century Quakers. Navigate to \\storageprojectsHISHIS-295-02Quaker_Network_Data and copy the quakers_nodelist and quakers_edgelist CSV files to your Desktop. Or, visit http:\/\/sarahjpurcell.sites.grinnell.edu\/digital_methods\/files\/<\/a> and download the same files.<\/p>\n

\n
Installing NetworkX<\/h5>\n

\"\"<\/a><\/p>\n

Python is already installed on the Library lab computers. To install NetworkX, navigate to C:\\Python27\\Scripts in File Explorer.<\/p>\n

\"\"<\/a><\/p>\n

Type cmd in the File Explorer navigation bar, and hit the Enter key. A command prompt window will open.<\/p>\n

\n
pip install networkx<\/pre>\n<\/blockquote>\n
Type pip install networkx in the command prompt window and hit the Enter key to install the NetworkX Python library.<\/p>\n
\"\"<\/a><\/p>\n
\n
pip install community<\/pre>\n<\/blockquote>\n
Type pip install community in the command prompt window and hit the Enter key to install the Community Python library.<\/p>\n
Close the command prompt window after the install has completed.<\/p>\n
If you want to install Python or NetworkX on your own computer, consult The Programming Historian\u2019s tutorial<\/a> for additional instructions.<\/p>\n
\n
Loading Network Data in Python<\/h5>\n
1-We will be working with the same nodelist and edgelist CSV files we downloaded at the start of this tutorial. Copy or move those files into your Python directory (C:Python\u2026) using File Explorer.<\/strong><\/p>\n
2-Click Start->All Programs->Python->IDLE (Python)<\/strong> to open Python\u2019s GUI coding interface.<\/p>\n
3-Start your Python file by importing the libraries and packages we\u2019ll need to use NetworkX.<\/p>\n
\n
import csv\r\nimport networkx as nx\r\nfrom operator import itemgetter\r\nimport community<\/pre>\n<\/blockquote>\n
4-CSV and itemgetter are built-in Python libraries, and we installed the NetworkX and Community libraries in the first part of this tutorial.<\/p>\n
5-Now we\u2019ll add code that tells Python to open the nodelist and edgelist CSV files and read the data contained in those files. Feel free to re-type this code, but you can also copy and paste from the online WordPress tutorial.<\/p>\n
\n
with open('quakers_nodelist.csv', 'r') as nodecsv: # Open the file\u00a0\u00a0\u00a0<\/em>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\r\n\u00a0\u00a0\u00a0 nodereader = csv.reader(nodecsv) # Read the csv\u00a0<\/em>\r\n\u00a0\u00a0\u00a0 # Retrieve the data (using Python list comprehension and list slicing to remove the header row)<\/em>\r\n\u00a0\u00a0\u00a0 nodes = [n for n in nodereader][1:]\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0 \u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\r\nnode_names = [n[0] for n in nodes] # Get a list of only the node names<\/em>\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\u00a0\r\nwith open('quakers_edgelist.csv', 'r') as edgecsv: # Open the file<\/em>\r\n\u00a0\u00a0\u00a0 edgereader = csv.reader(edgecsv) # Read the csv\u00a0\u00a0\u00a0\u00a0<\/em>\r\n\u00a0\u00a0\u00a0 edges = [tuple(e) for e in edgereader][1:] # Retrieve the data<\/em><\/pre>\n<\/blockquote>\n
\u00a06-<\/em>With open opens our CSV files in the read (\u2018r\u2019) mode and creates lists named nodecsv and edgecsv with that data. The additional indented lines read the CSV data, analyze the node names, retrieve the edge data, and remove the descriptive header information.<\/p>\n
7-From these commands, we have a list of node names and edge pairs which will be the foundation of our network analysis.<\/p>\n
8-Before moving forward, let\u2019s check to make sure our data has loaded correctly into Python.<\/p>\n
\n
print(len(node_names)) # prints the count of node names<\/em>\r\nprint(len(edges)) # prints the total count of edges<\/em><\/pre>\n<\/blockquote>\n
9-If you reopened the CSV files, you could confirm that Python had correctly loaded 119 node names (number of rows in the nodelist file) and 174 edges (number of rows in the edgelist file).<\/p>\n
\n
Creating a NetworkX Graph<\/h5>\n
10-Right now, we have a list of nodes<\/strong> (node_names) and a list of edges<\/strong> (edges) in Python. NetworkX will allow us to combine those lists in a network analysis, a Graph data object<\/strong> in NetworkX.<\/p>\n
11-Use G=nx.Graph() to initialize a Graph object. Right now the G graph object is empty.<\/p>\n