Twitter Germ Tracker to find ilness in locations in real time.

Twitter Data Helps Researcher Track Spread of Illness

Twitter is one of the most popular social media sites has helped people connect across the globe and has been vital in many situations. Now this micro blogging service has now helped Mr. Adam Sadilek, a researcher from University of Rochester to determine the spread of illness and other health risks.

Adam Sadilek has developed a mobile web app called Germ Tracker that will help us identify disease prone areas. For example, one can search for cold and flu affected areas using this app. The app works by analyzing Twitter data (tweets, locations, time, etc) in real time and alerts the areas which are affected most. According to the researcher and his team the app is able to analyze natural language for “meaningful trends” and discard the remaining.

Mr Sadilek claims that Twitter is a cheap, reasonable and dependable way for healthcare businesses to track the general health of a given area. On inspecting user messages in social media it gives a data of different population characteristics, which also includes the health aspect.

Twitter encourages everyone to express their ideas, thoughts, opinions and other random details of their lives. These updates vary in their importance and each individual update will not help but the aggregate of all the updates of these millions of users can give us important and valid information. Several studies relating to Twitter say that, even though Twitter is a micro blogging site it does give valuable insights into a population.

In this App, a broader range of public health applications for Twitter are considered. The recently introduced Ailment topic Aspect Model (ATAM)  is applied to more than over a one and half million health related tweets and the result was that it could discover mentions of a lot of ailments, allergies, obesity, insomnia and much more.

Extensions were introduced to incorporate prior knowledge into the model and apply other task such as tracking illness over time (syndromic surveillance), measuring behavioral risk factors, localizing illness by geographic region, analyzing symptoms and medication usage.

This shows quantitative correlations with public health data and qualitative evaluation of model output. Hence the results suggest that Twitter has huge potential in public health research in real time.

So, next time you pack your bags for a foreign trip, we’d recommend checking out this nifty app from Adam Sadilek to know what you are getting into and take necessary precautions.

Anand S

A techie who loves keeping track of the latest and the greatest in the industry.