Michael Jackson’s death created an emotional outpouring of unprecedented magnitude on Twitter. In this report, we examine 1,860,427 tweets about Jackson’s death in order to test various methods of sentiment analysis and gain insights into how people express emotion on Twitter.
Re-wrote Ruby Twitter scraper to be more efficient and designed and implemented a normalized database schema for storing tweets on Web Ecology Project’s server.
Original Twitter scraper prototype was coded in perl by Ethan Zuckerman (mentioned in a blog post on Apr 13, 2009). The script used Twitter’s URL-based API (since changed) to scrape tweets from a simple search query on a particular term like a hashtag. The script was ported to Ruby by Web Ecology Project member Dave Fisher, who also set up the initial database.
I re-wrote Dave’s code to make the scraper more efficient in how it handled the initial scraping of tweets and in writing to the database. I also designed a database schema for the tweets which organized the various metadata connected to each tweet in specific tables and columns that could be indexed for faster and easier queries across Web Ecology Project’s growing dataset.
My code was used to collect the tweets used in two studies I co-authored, Detecting Sadness in 140 Characters and Afghanistan and its Election on Twitter, and in another study authored by my Web Ecology Project colleagues, The Influentials.
Helped found the Web Ecology Project, an experimental community of social media researchers, in June 2009.
Details of Work
- Member of core strategy team
- Led business development trip in September 2009, presenting our research to a government agency and contractor in Washington, DC, and to a marketing agency in New York City
- Co-organized Web Ecology mailing list, transitioning organization from a for-profit venture to a distributed network of like-minded researchers
- Participated in all Web Ecology Camps: October 2009 (Boston), February 2010 (NYC), May 2010 (Boston), February 2011 (Boston & SF)
- Co-authored four studies using Twitter data
- Contributed to core Twitter scraping scripts and tweet database design
- Co-developed Web Ecology Project website
- Designed Adobe Illustrator template for all research reports and Apple Keynote template for all presentations
Related Publications and Presentations
- Graeff, E. 2011. Mapping an Arab Ecosystem (or two). Presented for the Mapping Media Ecosystems panel, MIT Center for Civic Media, Massachusetts Institute of Technology, Cambridge, MA, Nov 2.
- Lotan, G, Graeff, E, Ananny, M, Gaffney, D, Pearce, I & boyd, d. 2011. The Revolutions Were Tweeted: Information Flows during the 2011 Tunisian and Egyptian Revolutions. International Journal of Communication, 5.
- Graeff, E. 2010. Twitter and Democracy. Presented for the Internet and Society: Technology and Politics of Control class, Harvard University Extension School, Cambridge, MA, Apr 20.
- Graeff, E with Woodworth, S. 2009. Afghanistan and its Election on Twitter: The Macro Picture. Web Ecology Project.
- Kim, E, Gilbert, S, with Edwards, MJ & Graeff, E. 2009. Detecting Sadness in 140 Characters: Sentiment Analysis and Mourning Michael Jackson on Twitter. Web Ecology Project.
- (Various Authors). 2009. Reimagining Internet Studies. Web Ecology Project.