Conversationalist

Name: Sophie

Number of Teammates Desired: 2

Description of Idea:
The idea is to visualize potential conversations between 2 more or people. This application will tap into a number of social networking data to pull certain things such as user's likes, groups, interests, education, work, etc. and cross compare to find common points between people. We will find different areas of interest that are ranked according to the number of people interested in it, how often the topic comes up, etc. Then, based on the top few topics, we will create a mosaic of conversation topics.

For example, let's say 2 friends both are interested in Technology, Music (let's say they like Kaskade and Tiesto), and Movies (in that order). We will then create a mosaic with three blocks, Technology having the largest, music next, and finally the smallest being movies. Within those blocks, we will find images that correspond to good images that will represent potential conversation topics. For example, we can pull images off the tech crunch RSS feed about the potential iPad 3 release, and it will be represented as a tile in the mosaic.

This mosaic can be interactive, so when a user goes and clicks on the iPad 3 tile, it can enlarge and show the user the article from tech crunch. The interactions with this data visual can then be aggregated further to show general trends about conversations in general. Based on the tiles that people click on most common, we can have an aggregate "Technology" mosaic that represent the most commonly talked about tiles.

Description of Data:
This application will require taking data as input as well as output. For input, we will mostly rely on Facebook user data. However, we are really not limited in this way. We can just as easily connect to many different APIs: Foursquare, Twitter, etc. to gather more user info and therefore more to compare with between multiple users. For output, we will mostly use RSS feeds that contain information relating the user's common likes.

Additional Information: This is actually an area of research that I and one other person in this class are already involved with. So far, we haven't worked at all with data visualization or mosaics/tiles, but we have worked with Facebook API already to create conversation suggestions mainly RSS feeds for content generation. We have an online survey engine that gives out two conversation topics at a time to collect data on what conversation topics people would be interested in. If this visualization part of the application could be brought into play, this would end up being an incredibly cool project. Feel free to contact me (sophiez@andrew) to ask about what work we have already done with this project. I can give you a quick demo.