What's this site all about?

Following the Haiti earthquake on January 12th 2010 a variety of grass roots efforts sprang up in the software developer community aimed at providing help to the Haiti relief effort. Noone likes to sit around and do nothing in situations like this but we cant all leap on a plane and go and actually do something on the ground. We can give money to help the existing relief efforts and that is incredibly valuable but its still pretty passive so what else can we do?

One of the side effects of modern communication tools such as the internet, cell phones, smart phones, etc. is that many people now carry with them the means to communicate not only with their friends and family but with the world at large. Twitter has become one of these highly popular message streams and this was certainly the case following the Haiti earthquake. A huge number of Tweets asking for help, begging for aid and rescue crews, passing on information on missing persons, offering aid and assistance started coming out of Haiti and the affected area right after the earthquake hit. Its a simple thing to subscribe to streams of information like this using Twitter's search tools so many people were watching information flowing live from the affected area. This was quite different from watching on TV where the images are filtered through a camera lens or a correspondent with no way for us to reply. On Twitter you see the messages first hand from people trapped in the rubble asking for rescue, from people asking for information about their families. You could reply if you had something useful to say but what would you say, what could you do? Well, one thing you can do is try to help put that information into the hands of people who can can do something - the aid agencies and NGOs on the ground in Haiti.

Tweak The Tweet

This is where the Tweak The Tweet team at UC Boulder stepped up and started spreading the word about their work aimed at making Tweets more effective at conveying information. The basic idea is to rephrase the tweet and add in some simple hash tags to more specifically identify informative parts of the tweet, information on what, where, who, etc. This can make the tweet shorter and more compact (easier and faster to send) and much easier for other people (and other software) to parse and pull out the useful data. Here's a brief video from Denver TV that gives a good intro but some examples demonstrate it very effectively:

Sample Before & After Tweet Makeovers:

TWEET-BEFORE: roads from PAP to les Cayes are open migration from PAP to rural areas has begun

TWEET-AFTER: #haiti #open roads from #loc PAP to les Cayes are open #info migration from PAP to rural areas has begun

This tells the computer:

what = road
what about it = open
where = at PaP to les Cayes
what else: “open migration from PAP to rural areas has begun”

TWEET-BEFORE: Altagrace Pierre needs help at Delmas 14 House no. 14.

TWEET-AFTER: #haiti #name Altagrace Pierre #need help #loc Delmas 14 House no. 14.

This tells the computer:

what = need help
who = Altagrace Pierre
where = Delmas 14 House no. 14.

Parsing the tweaked tweets

With this structure in place one of the tasks becomes extracting this information from the Tweet stream, pulling out useful data, cutting down on the 'noise' and duplication and then placing this curated information back into the public domain for the NGOs and others to work with. That is what I'm experimenting with on this site: taking TtT-style tweets and parsing the information from them with the goal of then encoding this data using semantic web approaches (RDF, OWL, ontologies) so it can be integrated and augmented with other data out on the web. I presented some of these ideas at a seminar and I've put the slides up on slideshare so others can see what I'm thinking about. This is derived from our bioinformatics work with the National Center of Biomedical Ontology so it has some biology-related introductory material that sets the scene for a biology audience but might not be so understandable for non-biologists. The notes from the presentation are available on the main Slideshare site so viewing it there with the notes turned on might make more sense