This map displays all watches, warnings, and advisories issued by the National Weather Service. It's basically an animated version of the NWS WWA Display. I joined archival data from the Iowa Environmental Mesonet against mbostock's TopoJSON topology of US Census data. I don't update this anymore! Here's some data from back in the day: 2018, 2017, 2016.
The NWS issues polygonal storm-based warnings, but this map links alerts to entire counties. You'll sometimes see a county highlighted for a relatively small flood, or a Winter Storm Warning covering a whole county when only part of the county had a chance of snow (particularly out West). If the warning covers more than 10 kmĀ² of a county, the county is considered a part of the warning. I did this because I care about your data plan.
Start and end times for alerts are rounded down to the nearest 15-minute interval. Some alerts in the IEM data have bad end times (prior to their start times) due to weird data issues coming from NWS; those alerts are dropped.
Also, note that there's an entire class of marine warnings that are scrubbed out of this data.
Hey, that's not very nice. This just matches the NWS colors, though, so you'll have to ask them.
Because there are winter storms in Hawaii.
If you think you see a bug, tell me! But please read this first: the data from the NWS is imperfect, and some alerts in the IEM data are not terminated properly. Take, for example, this tropical storm watch that for some (but not all) counties hangs around for a couple of weeks in the IEM data while the rest of the original alert is expired. The problem is probably somewhere in NWS data, or possibly in IEM data, but regardless, I probably won't have the time to do point fixes on top of their data in the near future. I'm sorry!
I wanted a cheap proxy for "the weather was bad" to use for other projects. The animation was interesting to watch on its own, so I packaged it up.
The map was made with D3.js, TopoJSON, a few polyfills, and a little custom binary format for loading the warning data. The code was transpiled with Babel (hey there, IE11 users!).
The source for this page is on GitHub.
I'm on the bird. @jdhollen