Breaking

Post Top Ad

Your Ad Spot

Saturday, September 28, 2019

Sickweather Crowdsources Illness Reports for SickScore

Sickweather Crowdsources Illness Reports for SickScore

Depending on our plans for the day, we routinely check the weather using animated radar scans. Meteorologists predict the weather using algorithms that account for a variety of information sources, including data from remote instruments. Similar to weather reporting, Kansas City-based Sickweather‘s eponymous mobile app draws upon multiple data sources including social media to create localized illness predictions.

Sickweather has been archiving illness data since 2011. The company uses machine learning models to compare real-time sickness data reports with its archives in order to predict future illness rates. According to the company, the Sickweather app can predict illnesses up to 15 weeks in advance with 91% accuracy.

The Sickweather algorithms use several types of data: social listening and crowdsourcing, as well as hard data from the CDC, medication sales, and census demographic data. When people post information on social media such as Facebook or Twitter about being sick or family members who are sick, Sickweather captures the illness and location data. Individuals and Sickweather’s third-party partners can report anonymously to the company’s database regarding information about illness incidence. Sickweather uses CDC and other hard data to correlate and validate data gathered by social listening and direct reporting.

Sickweather’s app creates real-time sickness alerts on geographic maps. Each illness alert includes the number of reports in the area and the chance of illness for people in the area. A new Sickweather Apple Watch App displays a SickScore for your current location. The SickScore is an overall contagious disease risk rating. The SickScore app also displays the most common illnesses in the area and launches a handwashing timer.

Data quality is the Sickweather linchpin. The quality of the archived data used to train the machine learning algorithms determines the accuracy of predictions made by comparing real-time reports to the archives. Current illness report input validation is also crucial. When both sides of the data equation measure up, social media could demonstrate a tangible societal benefit beyond publishing selfies, emoji, and hashtags.







* Source

No comments:

Post a Comment

-

Post Top Ad