Database for PV panel location data

In the coming months we’re going to have more datasets of PV panels being published. This includes a big new global utility scale dataset coming from Descartes Labs and Smith School of Enterprise and the Environment (SSEE) at the University of Oxford.

Our goal is to have everything end up in OSM. But there seems to be a need to have a kind of open staging area where these datasets can be merged together, cleaned and de-duplicated (with existing panels pulled from OSM) before being prepared for submission to OSM. It might also be easier for users wanting to access PV data to use this (as it will be more comprehensive than OSM).

In essence all we’re talking about is GeoJSON data. So I don’t want to over complicate things. But I am wondering if there’s a platform which will help with the jobs of merging and cleaning.

Should we consider running our own OSM instance? This would give us access to the OSM tools, and ensure that everything we do is organised in a way that makes submission to OSM smooth.

@danstowell wdyt?

Hi @dct - yes, running an OSM instance could be a good way to provide “staging” data.

It’s not too difficult, it’s essentially a PostGres-and-Ruby website system, with clear instructions to spin one up.

Note that there are multiple components - for example, as well as the website and database, it might be a good idea also need to spin up a tileserver (which renders the data as a viewable “slippy map” so people can skim-view what’s in the data). The tileserver might not be needed though, depending how people are working with the data.

It could be done without a staging server, just by passing round geojson/gpx/osm files.

That sounds more than doable. I’m going to look into getting it running in the OCF GCP.