Segre A

Twitter is a free social networking and micro-blogging service that enables its millions of users to send and read each other’s ‘tweets’, or short messages limited to 140 characters. The service has more than 190 million registered users and processes about 55 million tweets per day. Despite a high level of chatter, the Twitter stream does contain useful information, and, because tweets are often sent from handheld platforms on location, they convey more immediacy than other social networking systems.

Objective

This paper describes a system that uses Twitter to estimate influenza-like illness levels by geographic region.

The rise and associated risks of using the internet to find sexual partners among men who have sex with men (MSM) has been noted by many researchers.1,2 The anonymity and relative ease of finding partners on the internet has facilitated casual sexual encounters that can encompass a variety of unsafe sexual practices, from anonymous partners to ‘Party and Play’ activities (PNP), slang for illegal drug use, unprotected sex, group sex and so on. These anonymous sexual encounters make it more difficult for public health officials to notify exposed partners. In addition, detailed data regarding risk behaviors are generally obtained via conventional survey techniques, which are expensive to conduct. Thus, a general method of empirically deriving large scale, location-specific behavioral data could be immensely useful in understanding or anticipating STI outbreaks. Craigslist is a website specializing in online classified advertisements around the world. Our hypothesis is that Craigslist contains rich behavioral data regarding MSM communities and that such information can function as proxy for external prevalence rates for diseases (that is, HIV/AIDS).

Objective

This paper describes a novel method of obtaining large scale, geographically diverse behavioral data about Men who use the Internet to seek Sex with Men (MSM) by examining anonymous Craigslist message posts to predict HIV/AIDS.

Prediction markets are a type of futures market in which users trade shares that pay off if the event to which they are connected occurs. They are used to aggregate knowledge on a large scale, as the prices of the various contracts can be interpreted as probabilities of their events. Since 2006, our group has been using prediction markets and testing their utility in predicting the spread and impact of diseases, including seasonal influenza, syphilis, and others on a market called the Iowa Electronic Health Markets (IEhM), found at http://iehm.uiowa.edu. For example, in 2009, a series of markets were run on novel influenza A (H1N1), which showed success in predicting the extent and duration of the outbreak.1 We currently plan to move into a new phase of development that will allow the community of users to submit proposals for new prediction markets, which will then be approved by site editors and referees. We call the new system Samos.

Objective

This poster presents a software system to provide a community-driven, user-generated, low-overhead, web-based prediction market system called Samos.