Class Projects - Fall 2013

TwConnect

TwConnect helps Twitter users get to know each other more quickly by finding interests they have in common.

TxConnect is a proof of concept technology that helps Twitter users get to know each other more quickly by finding interests they have in common.
What it does

Finding common ground with new acquaintances— whether they’re professional, personal, or potentially romantic — can be difficult and somewhat stressful. TwConnect seeks to make the getting-to-know-you process easier by allowing users to quickly find topics of mutual interest based on each person’s Twitter histories. Users enter their own and an acquaintance’s handle in to TwConnect and the system returns topics (sports, technology, politics, etc.) and connections they have in common.


How it works
When a user enters two Twitter handles in to TwConnect the system connects to the TwitterAPI to look at profile information and tweet history. Next, the system uses the AlchemyAPI to extract key words and entities from past tweets, taking in to account word frequency and outliners shared by both subjects. Consolidated keyword-related tweets are then sent to AlchemyAPI to get categories for key words and cross-matched against the categories in a proprietary database. Combining categories and other profile information, all data is organized and sent back to the user.
Next Steps
  • Current categories are limited to sports, politics, technology and entertainment. A richer classification system and subdivisions on each category would be an improvement.
  • Adding a sentiment analysis to the system, which would show the attitudes of the user towards each topic, instead of just showing the frequency.
  • Though named TwConnect, the service might be more compatible with other social network like Facebook.

TwConnect architecture.

TwConnect screenshot.

Connect

Student Team: Cameron Johnson, Garima Jain, Zhenyu Zhan

Faculty Guidance: Rich Gordon and Larry Birnbaum