University of Maryland

TUTORIAL: Social Media Analytics During Crises 2016

TUTORIAL: Social Media Analytics During Crises

Thursday, May 26, 2016

Workshop on Social Media during Crises

A tutorial during the
33rd Human-Computer Interaction Lab Symposium
University of Maryland

Overview

This tutorial will build practical experience in using Python + Jupyter Notebooks to analyze and discover insights from social media during times of crisis and social unrest. We demonstrate how temporal, network, sentiment, and geographic analyses on Twitter can aid in understanding and enhance storytelling of contentious events. Examples of events we might cover include protests in Ferguson, MO, the Boston Marathon Bombing, and the Charlie Hebdo Attack. Demonstrations will include hands-on exercises in extracting tweets by location, sentiment analysis, network analysis to visualize groups taking part in the discussion, and detecting high-impact moments in the data. Most of the work will be performed in the Jupyter notebook framework to aid in repeatable research and support dissemination of results to others.

Prerequisites

Tutorial participants are expected to have some prior experience in programming. Proficiency in Python is preferred but not essential as Python is a straightforward language to learn given prior experience with C/C++, Java, Perl, etc.

Tutorial content will be built on the IPython/Jupyter notebook framework, which does not come standard on most platforms and is generally installed via the command line, so a familiarity with console applications is also preferred.

Organizers

Questions: Please contact Cody Buntain (cbuntain@cs.umd.edu)

Agenda (Subject to Change)

The precise timing is not set yet but it will likely follow this format:

  • 08:15am – Symposium Registration, Breakfast
  • 09:00am – Symposium Plenary Talks (more information)
  • 1:00pm-1:15pm – Tutorial Introduction
  • 1:15pm-3:00pm – Tutorial: Session I
    • Topic 1: Introducing the Jupyter Notebook
    • Topic 2: Data sources and collection
    • Topic 3: Parsing Twitter data
    • Topic 4: Simple frequency analysis
  • 3:00pm-3:20pm Coffee Break
  • 3:20pm-4:30pm – Tutorial: Session II
    • Topic 5: Geographic information systems
    • Topic 6: Sentiment analysis
    • Topic 7: Topic modeling
    • Topic 8: Network analysis
  • 4:30pm-4:45pm – Tutorial Conclusion
  • 05:00pm – Symposium Demos, Posters, Reception
  • 06:30pm – Symposium Ends