Monday’s short but devastating storm in Chicago reminded me again of a product based on open data that I wish someone would make: a Severe Weather Community Center. I think there would be two parts to this: historical knowledge/ learnings/ upshots and a ready-to-go infrastructure when things go south.
I picture the historical info to be incident-based (“Monday Morning Rush Hour, July 11, 2011”). This info would be pulled from a number of sources:
- National Weather Service, showing the paths of the storm down to the lowest levels, showing how they swooped through the city
- City of Chicago Emergency Tree Service Requests: showing the areas where the storm knocked down trees. This data is not yet available, so I made a suggestion via this form
- Department of Streets and Sanitation data: they love to talk about how many trucks have been running for how many hours, when they’re ready to get to the side streets, etc. This could be in the form of data, or just via reporting. Next time Tom Byrne is at the podium being all command center-y, it would be cool if someone added his truck force data into this system
- Flickr and other services: showing pictures of storm damage. I’ve got a goodly collection of this stuff myself.
- Chicago Transit Authority: showing number and length of service outages
- Utility data: showing number and length of service outages. This one is a toughie— very difficult to get this information out of them, but it is clearly available. Also, they keep talking about the Smart Grid, and how it is going to make things better. Let’s go ahead and get smart. UPDATE, October 10, 2011: Here’s a starter ComEd dataset published and analyzed by the Chicago Tribune
- Traffic data: showing deviation of traffic times from the norm
- Moar, moar, moar. There’s lots of data that would be useful to a system like this. Let me know your ideas in the comments
After a while, I’m guessing some patterns would emerge. We would see that a late Spring storm with winds of 50 miles an hour striking in mid-afternoon knocks down 700 trees, affects the trains for about 45 minutes, and makes suburbanites about an hour and a half late getting home. A winter storm coming from the northwest in early February lays 15 inches of snow and it takes 14 hours after the snow stopped before trucks can get to the side streets. We could use this more granular shorthand when we’re preparing for a storm.
Which brings me to the ready-to-go infrastructure. After a storm strikes, there’s a lot of confusion, lack of essential services, and plethora of people looking to contribute. The way Craigslist was adopted after Hurricane Katrina is the classic example of people using existing technology to meet their needs in an emergency. The problem is that, 6 years after the storm, it doesn’t seem like there is a unified solution to the recurring problem of natural disasters and weather emergencies. (I might be wrong— I haven’t done an exhaustive search. Let me know what’s out there).
The thing that comes closest is Ushahidi, the excellent platform for information collection, visualization and interactive mapping. I got some experience with this system when I helped the Chicago Tribune News Apps Team manage snow reports during the blizzard of February 2011 in Chicago in their Ask for help, lend a hand: Blizzard 2011 project. That system was a pretty good for receiving inbound requests (though the geocoding was fragile), but there was no top-down aspect to the communication (“check when ComEd says the power on your house will be turned back on”). We need both in order to have an effective system.
With the data feeds listed above, especially the power outage info and the times that city service, the system could be a good clearinghouse. This is great for people who have electricity and are able to access the Internet. But we know that the communications infrastructure can perform poorly when it is taxed in an emergency. I saw from my own experience with Twitter on Monday, trying to update CTATweet, that the service was unavailable right when I needed it most.
What I’d love to see is a system of real-world places where this information could be added to and disseminated, regardless of the status of the electrical grid. This could work like the old newspaper offices. Seeing this post about the Boston Globe’s intown system, where they posted updates about current events that were being published in the paper right on their HQ window, made me smile. Papers really are essential, especially in times of need.
Update: Shortly after I posted this, I got this notice in my mailbox:
Definitely would be nice to whack tree trimming schedules against tree down reports to see if there is a correlation between tree maintenance and tree limb falls.