You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I've been messing around with an improved flow utilizing a mapping between the Census API and TIGER API. It allows for easy and modular support of census defined geographies. Using this mapping, SQL WHERE queries can be constructed using a defined hierarchy to pull geometry data in a single TIGER query (as opposed to pulling a geometry and using the bounds for another query).
I've gotten it to work with the ACS using almost all of the defined whole geographies (ignoring (or part) geographies for now and failing on congressional districts and zip code tabulation areas). See the latest build and tests for coverage. The key piece of this is located in a configuration file where the mapping is defined.
A few limitations (at the moment):
Non-defined census geography hierarchies - such as pulling tract from combined statistical areas - need to be identified and handled separately
(or part) geographies (might?) require identifying multiple geometries, querying multiple layers, and taking an intersection of the two
However, I believe it's doable to get full support from both census defined and non-defined geographies with a little bit of work. This applies to both the ACS and Decennial products.
BIG NOTE: Getting this functional did require some reworking of the modules and I admittedly got very carried away. I made a handful of changes while I was reworking everything. It is very different from the original. Seeing if this was a worthwhile endeavor beforehand and consulting with the group would have been the smarter play.
Integrating this in would require a handful of work like testing, documentation, and adapting for backwards compatibility. Is this something that would be worth to start integrating into cenpy and continuing development in this repository? Perhaps 2.0?
The text was updated successfully, but these errors were encountered:
See the development branch here: https://github.com/ronnie-llamado/cenpy/tree/development
I've been messing around with an improved flow utilizing a mapping between the Census API and TIGER API. It allows for easy and modular support of census defined geographies. Using this mapping, SQL WHERE queries can be constructed using a defined hierarchy to pull geometry data in a single TIGER query (as opposed to pulling a geometry and using the bounds for another query).
I've gotten it to work with the ACS using almost all of the defined whole geographies (ignoring
(or part)
geographies for now and failing oncongressional districts
andzip code tabulation areas
). See the latest build and tests for coverage. The key piece of this is located in a configuration file where the mapping is defined.A few limitations (at the moment):
tract
fromcombined statistical areas
- need to be identified and handled separately(or part)
geographies (might?) require identifying multiple geometries, querying multiple layers, and taking an intersection of the twoHowever, I believe it's doable to get full support from both census defined and non-defined geographies with a little bit of work. This applies to both the ACS and Decennial products.
Integrating this in would require a handful of work like testing, documentation, and adapting for backwards compatibility. Is this something that would be worth to start integrating into
cenpy
and continuing development in this repository? Perhaps 2.0?The text was updated successfully, but these errors were encountered: