Request for a detailed repo structure

[dhaval-shukla]

The suggestion is to have a more detailed folder structure or attributes for the maps here.

Suggested attributes:
Admin level, name of largest admin level, year of data, boundary code.

• Admin level is 0, 1, 2, 3, 4, 5, 6...(0 country, 1 state, 2, district, 3 taluka, 4 village, 5 ward, 6+ assembly/parliamentary constituencies etc)
• Name of largest admin level: If it is just a state or district map, then we would need to know the name of that
• year of data- self explanatory
• boundary code- which boundary does it follow (explanation below)

Reasons:

• There are tons of requests for boundaries of different admin levels all the time in Datameet google group, and the data gets unorganized
  -Ward Boundaries may get updated during non census years. Ahmedabad itself has had 2 updates since last census- from 64 wards to 57 to 48 wards.
• District boundaries get updated during non census years
  https://blog.socialcops.com/intelligence/data-stories/boundary-changes-districts-in-india/
  According to this, 44 new districts have been created since the last census (2011) for 12 different states.
  -Boundaries are not the same in all maps

Example:
Recently, I needed a district level map of Gujarat for a publication. Gujarat's districts had been updated in 2013, but the ones here showed 2011. I found some district maps from other sources, but they were wrong. Finally, someone had nearly accurate maps on the group, and I used them for this.

This is why I believe it is important to have state and district level maps organised by district/state as well as year.

Similarly, if we know which boundary a state or district map follows (GADM, OSM, Natural earth etc) then comparing data between two geographical entities would be easier. This is more for the visualization purpose

For example, if I want to compare 3 states on district wise population density and show that on a larger map of India, I need to make sure that all 3 district maps follow the same boundary/map source.

[answerquest]

Some ideas:

• Having a circa year (like: this is the boundary as of..) in the filename or nearby may help. For Pune wards I'm keeping years (pre-2012, 2012-17, 2017-onwards) in the filenames.
• I think the folders, files are organised enough right now. I wouldn't want to put just 0,1,2 etc and require people to look up to know what's district, what's state etc. And we don't have some major programming requirements either to put a 0 / 1 / 2 etc in the main names.
• Village boundaries might be better off in their own repo's.
• Sub-dividing into multiple repo's along admin levels or regions may be one expansion strategy as this grows, as typically someone wanting, say, districts may not want to download another few 100 mb's of other boundaries.

Annnnd:

• I think if we want it badly enough then let's just fork, re-organise, pull-request or just post back here, help do instead of asking/demanding to do.