Back to top

Flu affects about 5-20% of US population each year!

On average, tens of thousands of individuals die from Influenza in the United States each year according to the CDC. Flu or influenza is a respiratory illness caused by influenza viruses that infect the nose, throat and lungs. It can cause mild to severe illness and even lead to death at times.

Flu can spread through several ways but it is most commonly spread through air or by touching flu viruses. It is said that prevention is better than cure, so what if flu can be prevented by analysis and visualization of factors causing and influencing it?

Our application uses interactive visualization to predict disease occurrence based on climate, travel and geospatial data. This real time application will help prevent flu by advanced disease prediction. Flu Map uses IBM Bluemix tools like Node red and sentiment analysis to analyze social media (tweets) in real time.

Social Impact

Flu surveillance systems can make enormous social impact. According to the CDC, between 1976 and 2006, estimates of flu-associated deaths in the United States range from a low of about 3,000 to a high of about 49,000 people. Seniors, young children, pregnant women, and people with certain health conditions can be more susceptible to the flu virus. Our project aims to help people who are at high risk for serious flu complications by keeping them informed when there is high flu activity at places where they live or plan to go. Furthermore, our goal is to make the information as accessible as possible so they can stay constantly vigilant.

Back to top

Watch It Work

Below you see the estimated number of people who visited the doctor with flu-like symptoms on the given day for each county in the United States.

For each day today or before, the estimate is based on tweets coming from the area specifically about the flu, combined with the Google Flu Trends estimate for this week. For each day after today, the numbers are based on a regression model which has been derived through a machine learning algorithm which utilizes a simple model of disease spread, taking into account factors such as population density and travel between counties. For more information, go to the section How it works

Back to top


Want to use this data in your own app? It's simple!

Simply send a GET request to<SS><CCC>, where <SS> is the 2 digit state FIPS code and <CCC> is the three digit county FIPS code of the county you're interested in. The U.S. government keeps a mapping here to help you find your counties. They also provide an api, which takes latitude and longitude and converts it to the full 15 character FIPS code, the first 5 of which is what you'll want.

As an example, the FIPS code of Marion County, Indiana (which contains the wonderful Indianapolis), is 18097.

Back to top

How It Works

Our application incorporates multiple publically avaiable civic data sets and utilizes various technologies to create a prediction of how flu might spread over the next 7 days by incorporating various social, weather, travel, and open data sources. We leverage a stack that is flexible on BlueMix and have built a machine learning algorithm in Node Red which works over time to tune the prediction. Our findings are then presented on a simple to use map shown above.

Bonus: As we built this application we contributed to a Node Red / BlueMix patch which enabled Node Red to fetch files from HDFS of size greater than 5Kb. C.f. the fix on GitHub. Luckily it was fixed before the deadline!

Application Node Red / BlueMix Flow

Download More About How This Works

Back to top

Our Team

Our team is made up of a diverse set of backgrounds led by Perscio and supported by Purdue University students

Matt Gilliam Alec McGail Brian Norris Bob Boehnlein Dr. John Springer Anuja Rayarikar Tao Wei Kao Yuan-Hsin