Included obm_built_area_size column in obm_tiles

With this command the obm_built_area_size column is added to obm_tiles. It contains built up areas based on the portion of the OBM building footprints that intersect with a tile. also projected to Albers Equal Area.

\approve @ds @fd

added 1 commit

• a3dae161 - Changed built_area_size_obm column name and doc

Compare with previous version

For me, the overall question here is, if the OBM Buildings size should be stored in the main obm_tiles table, or just as a different source in the obm_completeness_assessments.

In my opinion this is information that has to do with the tile itself rather than a completeness assessment. The sources are oriented to external datasets to perform the assessments

From my perspective, OpenStreetMap (where these buildings are coming from) is an external dataset, too. The benefit that I'd see to have it in the obm_completeness_assessments table, is that we can also document the date of calculating the size derived from the buildings.

That would imply that we compare OSM with OSM as well?

I don't understand. What do you mean?

What this is about is to have settlement data (stored in obm_completeness_assessments) to be compared and assess its completeness with the state of the tile which is given by the OBM/OSM built area. So having OBM/OSM built areas in obm_completeness_assessments doesn't make sense to me and would feel like compare OBM to itself.

When I said OpenStreetMap above, I referred to OBM Buildings, as this is where the data is coming from, and I wasn't suggesting to compare OSM with OBM Buildings. Agreed that this wouldn't make sense at all.

Datasets saved in obm_completeness_assessments can be compared the same way if they are in the other table. By adding it to the obm_tiles we just narrow down a lot of good flexibility that we could well need later on and we build a more monolithic system that will be harder to manage.

Besides the increased flexibility and following a recommended modular approach, I think it would be nicer for visualization purposes, too, having the OBM built area size data in the obm_completeness_assessments table, as we could use the same style and tool to visualize all kind of datasets that are being added.

I understand better now. Then, I think we can transform this new field into a source_id == 0 in obm_completeness_assessments and do the calculations comparing other sources with source 0. What do you think?

@ds What is your opinion about having the OBM built areas to be another source in obm_completeness_assessments instead of having it in obm_tiles?

@nicgar: I always thought they should be in obm_completeness_assessmentas a source for any further assessment. However, as discussed initially,obm_tilesshould reflect our best knowledge about the tile, i.e. built area and its completeness (being manual or automatic, whatever is the best estimate). Either we have the polygon of built area inobm_tilesor we refer to it in theobm_completeness_assessmentor inobm_sources`. I really don't care where the data is stored as long we can easily disseminate and use the best available information per tile.

I will close this MR

closed