Two new services came to my attention this weekend, and while they aren't technically related, they cover subjects often found together: weather and traffic. Current traffic reports and weather forecasts are staples of commute-hour news reports in pretty much every country I've ever visited, and (at least in the US) some radio stations compete on the basis of just how often per hour they can squeeze in updates. What makes these two new services interesting is that they flip the format: as this post's title subtly suggests, it's the weather that's current and the traffic that's the forecast.
The Google Weather Map is a YAGMH (Yet Another Google Map Hack), a category which seems to grow daily -- and will undoubtedly be soon joined by the YAYMH for the Yahoo! maps. The GWM combines the map with weather data pulled from two sites: Weather Underground, which pulls current data from the Automated Surface Observation System (ASOS), largely located at airports around the US and territories, and (less frequently) from weather stations around the world; and Weather Bug, which has about 8,000 tracking stations in schools, TV stations and residences around the US. If you like images to match to data, the GWM also has links to weather webcams.
As Google Map hacks go, it's not too bad, but it is somewhat unstable. The hosting service seems to fail frequently; if you get an "unable to link to script server" message, give it a few minutes and try again. It's also partially broken in Safari (the data links come up blank) and MSIE (data links are doubled), but works fine in Firefox. The International support is spotty, with some locations coming up blank, and others (such as Paris) showing weather flags but no map data. Odd. Still, as a proof-of-concept, it's pretty cool -- and I particularly like the distributed, participatory nature of the data services.
Real-time weather services are familiar, if not overly common. Much more novel is the Seven-Day Traffic Forecast from "BeatTheTraffic.com." As described by New Scientist, the system:
...combines real-time traffic density and speed with historical trends on major routes. [...]The level of congestion depends largely on the time of day and day of week, and often patterns are repeated through the seasons. “With the proper tools, you can predict quite a bit of this,” he says. But the system can also use past data on the effect of accidents to predict what will happen after a new one occurs.
The system takes real-time sensor information from major routes and junctions as its inputs. In addition to this it uses alerts on traffic related weather conditions, road incident alerts, road works and even some calendar events such as sporting events, in order to make its predictions.
The service is being rolled out in various locations, but requires either an annual BeatTheTraffic subscription or a local TV station showing the data during the news. Fortunately, while the application is novel, the underlying ideas are not -- we will almost certainly see other locations pulling this data together (and probably combining it with Google Maps) in the coming days and weeks. It would be particularly interesting to link this to public transit information and planning maps.
Comments (2)
Development server is having difficulty keeping up with high levels of site traffic (much higher than I anticipated). Planning to more stable production environment to handle load. Migration to Google Maps version 10 should fix some of the reported issues with Safari and Internet Explorer.
http://www.weatherstationmaps.info
Posted by dave | July 3, 2005 3:25 PM
Posted on July 3, 2005 15:25
Thanks, Dave.
You should take the higher-than-expected traffic as a sign you've hit upon a very good idea.
Posted by Jamais Cascio | July 3, 2005 4:57 PM
Posted on July 3, 2005 16:57