Metafeed

We have been primed to ingest streams of content through "feeds" (whether social, rss, or just news), with information typically presented in staccato snippets that glimpse at the contents. While we're very happy drinking from the firehose, there may perhaps be a more palatable way to present this stream, one that not only could see past marketing-driven headlines and stock photos, but also see how one article relates to others. This app uses text mining techniques (clustering; distance between documents) to display content from a higher-level perspective, presenting them as a meta-level visualization of your feed. The graphic representation provides an alternative, faster, method of looking at large streams of data, helping you filter your list quickly, so you can spend more time reading content you actually value.

Metafeed began as a final project for Metis Ruby on Rails intensive, and was then taken on as a longer-term research project – looking at it from the perspective of HCI and text mining. A short paper was submitted for the Applied Human Computer Interaction course at AUT and is currently being worked on under the guidance of AUT professor, specializing in machine learning.

Project Status: Personal project. Currently in-progress.

Team & Role: Individual project. Doing everything (concept; design; implementation; etc.).