Outbreaks of infectious disease are common in all parts of the world, and knowing about them as they happen is crucial both for governments and for individuals living in the affected areas. Google Flu Trends has done an amazing job of plotting flu outbreaks in realtime across the U.S. But while Google's data is organized at the state level, outbreaks are often of a much more local than that, affecting city populations, or even ultra-local neighborhood populations. How do we detect for such local outbreaks as they emerge? While it isn't necessarily important to you that there's an outbreak of flu in your state, you certainly want to know if there's an outbreak in your city. What to do?
Enter SickCity 1.0, the newest app launched at DIYcity. SickCity watches users' status messages on social networks like Twitter and Facebook, monitoring these feeds in realtime for keywords like "flu", "chicken pox", "fever", etc. It then plots the results on a chart showing 30-day trends for those words in your city. The result? A simple snapshot showing you whether conversation about these things is on the rise, on the fall, or not moving. And that may give you some early indication of whether an outbreak of a particular disease is in the works. After all, when people who use Twitter or Facebook get sick, what's the first thing they do? Tell their friends about it, of course.
SickCity is a bottom-up, realtime alert system for outbreaks of infectious disease in your city, configurable by anyone and costing essentially nothing to run.
Okay, but does this really work?
Well, we're not saying absolutely one way or another at this point. This is a prototype, and it just launched. A bit of study and reflection is in order first, along with some refinement. But there's no reason to think it shouldn't, in areas where social app use is high. And the higher the usage, the more likely the results are to be accurate. And as we refine the filters in subsequent versions, the results are likely to improve as well.
[UPDATE 3/23/09: the team building SickCity has been joined by a statistician who is crafting algorithms to improve SickCity's accuracy in determining a city's "sickness". we have also been contacted by a public health group who wants to correlate our findings with public records on diseases in cities.]
Go to http://sickcity.org, navigate to your city, and view the results.
This initial release is hard-coded for the beta cities we have chosen. Some mechanism will be developed quickly to allow for more cities to be added. Details to follow.
Next things to do for SickCity include:
- eliminating false positives and generally improving filtering algorithms for better accuracy
- allowing people to drill down to their neighborhood to see what the trends are for specific disease keywords
- adding an email alert system for people who want notification of certain keyword trends
if you have other thoughts for how to improve SickCity, please start a thread in the Discussions group.
The mission of DIYcity is to offer free, open source tools that communities all over the world can use to make their local areas better. The idea is to build a generalized toolset that can be applied roughly to any local community, then let individual communities tweak those tools as necessary to suit their needs. Those communities then commit their changes back to the DIYcity code repository so that other communities can make use of those modifications as well.
In this way, DIYcity generates a toolset for communities everywhere, complete with specific modules to make the tools work better in varying local circumstances.
Sickcity is the second application to be built by the DIYcity community with the aim of fulfilling this mission. (see first DIYcity app here).
If you like this idea, please get involved by joining DIYcity and participating.
Version 1 of this app was conceived, designed and built by:
Questions, comments, ideas? Please contact us at firstname.lastname@example.org. Or just join the Discussions group and say hi.