Online Program

Saturday, October 22
Knowledge
Community
Influence
Sat, Oct 22, 11:00 AM - 12:00 PM
Salon 2
Speed Session 4

Social Listening on Twitter: Understanding the Data Science Conversation (303239)

*Adrianna Jelesnianska, Civis Analytics 

Keywords: Twitter, Social Media, Text Analytics, Natural Language Processing

Twitter is a notoriously noisy data source, but its real-time availability and scale make it worth the time and effort to search for a signal. Civis has developed a social listening tool to understand the underlying trends and conversations for a specific issue under discussion on Twitter. This talk will present an analysis of tweets on the topic of Data Science and include a brief introduction to the tools and methodologies that allow users of our system to form a cohesive story around the issue. Specific software tools include python libraries such as gensim, nltk and a java library that helps us detect the language of a tweet. The methodologies discussed will be topic modeling, network analysis and sentiment analysis. As we will demonstrate during the session, our system allows users to find the answers to many questions that are not easily resolved by looking at a typical social newsfeed. These include how many tweets (if any) discuss women in data science, and how the topics have changed over time in data science.