Closed spatialaustin closed 8 years ago
I'm doing some cleaning and will attempt to match up with the city's address points layer. It looks like they're coming from a free-text field; there's a lot of variation in formatting.
This is for the historical dataset. We should be able to work within API quotas (google, bing, etc) for daily permit pulls.
i've added lat/lon to the permit reports.
now that the backlog has been cut down, a geocoding library could be used to handle daily report dumps. e.g. https://pypi.python.org/pypi/geocoder
the permit-pull + geocoding script is close. i need to relearn how to write a dictreader instance to file and that's about it: https://github.com/open-austin/construction-permits/blob/more-geocoding/dank-eshet.py
there's going to be another backlog of permits to geocode. boo-hoo.
this process needs to be repeated. holding off until geocoding chron job is set up.
Things I've learned using geopy:
location = geolocator.geocode(address, components={'locality': 'austin'})
(location.latitude == 30.267153) and (location.longitude == -97.7430608)
Good luck!
all very great tips, thanks!
backlog of permits to 1980 geocoded and committed.
i also renamed all of the pre-chron job cvs to match the YYYY-MM-DD convention. so that's good.
The permit addresses need to be geocoded - they do not have lat/lon info. There are ~180,000 unique addresses