Blog Post

Reflections on Twitter Analysis

The goal of our project is to perform analysis of “black twitter”, or more generally to examine racism and race issues as related to African Americans on Twitter. In order to do so, we compiled a list of keywords that are commonly used or associated with these issues, and will use these to compile a set of tweets containing any of these key words or phrases. After doing so, we want to perform different forms of analysis in order discern interesting data. An example of this is a heatmap: we wish to see where in the world these discussions are taking place, and what that tells us about these issues.

My role in the project up to this point has been primarily technical. I have been working to develop a tool in Python which will download tweets based on keywords given in a csv file, and after downloading them I have been developing a tool to parse out the relevant information, such as the text of the tweets or the location data. I have some experience working with the Twitter API already, and besides that I am a computer science major so I bring strong technical skills to the group. I plan to use my strengths in order to collect the pertinent data, and then process it such that it is in a form that everyone in the group, myself included, can then create visualizations and analyses of it.

One of the primary technical components is Twitter itself obviously, and along with that there are my tools with download and process the tweets. We are considering methods of mapping tweet location data, and plan to use Voyant for text analysis. I think that although we are interested in studying the discussions against racial inequality that appear on Twitter, that often times Twitter itself can be a platform for hate and racism. I think that Twitter, and nearly any other social media platform, provides a way for like-minded people to find each other and unfortunately some of these groups are hateful. In spite of this, there are still important conversations on Twitter against racism, sexism, etc. and we wish to bring these conversations to light. 



I really like how your comment about how Twitter is often (and notoriously) used for hate. We need to be mindful that we expose and not inadvertenly propograte or confuse data with hate. For example when we map locations out, how and do we want to avoid mapping out 'hateful' responses or trolls?


Jamie, you bring up a great point about hateful responses and trolling.