jtkovacs / groupby

uwseds-group-group-by created by GitHub Classroom

ANALYSIS OF SOCIAL MEDIA DATA

Project Description

Browsing, interacting, and creating content on social media has become a way of life for many people. Although some or all of one's social media data can often be downloaded, there appears to be a dearth of tools to facilitate analysis of that data--for example, NameGenWeb was created by the Oxford Internet Institute to assist Facebook users in exporting their networks, but the tool is no longer maintained or usable.

Because we believe that social media users--not only marketers--deserve to understand their social media behavior, we are attempting a tool project that will enable users to gain very simple insights into their social media use. Currently, our tool is capable of ingesting datasets from Facebook, LinkedIn, Twitter, and Google Calendar. In addition to providing some summary of activities on each platform, we anticipate enabling users to explore several questions that depend on integration of data sources:

• Do their levels of activity across different social media platforms rise and fall together, or do they cycle between one platform or another?
• What times of day do they tend to be most active on social media?
• How does their social media activity compare with their other activities, for example, how often does social media activity overlap with scheduled events?

Getting Started

Install

Installation makes our code available as a command line tool as well as a Python module that can be imported and used within your own code:

• Clone the github repo: https://github.com/UWSEDS-aut17/groupby.git
• From the command line, navigate to the repo and run these commands:
  • pip install .
  • pip install -r requirements.txt

Uninstall

• pip uninstall groupby

Testing

• python setup.py test

Obtain your data

To use this tool, you must download and unzip at least one of these datasets from your personal accounts (note that there can be a delay of a day or so, depending on the platform):

• Facebook - Official instructions
• Twitter - Official instructions
• LinkedIn - Official instructions
• Google Calendar - Official instructions

Use the tool (command line)

Option	Argument	Example Argument
-T, --twitter	Path to Twitter data directory	~/Downloads/Twitter
-F, --facebook	Path to Facebook data directory	../backups/Facebook12-20-2014
-L, --linkedin	Path to LinkedIn data directory	data
-C, --calendar	Path to Calendar data file	~/Documents/mycal.ics
-O, --output	Path to desired report output location	~

The simplest use case is to analyze a single social media dataset:

groupby -T path/Twitter_directory -O ~/Desktop

You can analyze multiple social media datasets simultaneously but separately:

groupby -T path/Twitter_directory -F path/Facebook_directory -O ~/Desktop

Finally, you can analyze all social media datasets in conjunction with calendar data:

groupby -T path/Twitter_directory -F path/Facebook_directory -C path/Google_Calendar_directory -L path/Twitter_directory -O ~/Desktop