This Dutch mash-up is based on PM10 data (particulate matter <10mg) from the Dutch national institute for environment RIVM. It takes the data and builds a smog forecast as well as historic record for each of the RIVM's sensors. Data is stored in Google App Engine.

With Vervuilingsalarm you can set an alert if the forecast for a certain sensor reaches a level you indicated. Alerts are distributed to you via Twitter at @rivmalarm.
Sensor-data is also fed into Pachube. Pachube is also used as primary source for the sensor's GPS coordinates.

Both coordinates and complete dataset for any RIVM sensor additionally is made available in JSON through a URL structure like: http://vervuilingsalarm.nl/station/133/json/ (where 133 is the sensor used in this example)

Contacts: Ton Zijlstra, James Burke, Alper Çugun.