Improve data extraction based on survey categories #2959
Conversation
github-actions
bot
added
enhancement
|
Doesn't it make sense to keep OSM in the name, as they are forms specifically tailored for OSM data collection? Other forms we advertise may not be OSM related at all 😄 |
There was a problem hiding this comment.
I assume Rob (the previous tech lead) spent some time working out the specific tags in OSM that apply to each category here. Did they not work as intended?
There was a problem hiding this comment.
I am not sure if this has something to do with raw data api, but in most cases the original JSON data models were returning almost all buildings every time regardless of the form type chosen. Upon some testing, I found the new json data models I wrote worked much better in capturing data.
| "amenity": [] | ||
| }, | ||
| { | ||
| "shop": [] |
There was a problem hiding this comment.
For example, I would say a shop is probably an amenity
| "amenity": [] | ||
| }, | ||
| { | ||
| "convenience": [] |
There was a problem hiding this comment.
It might be a bit counter intuitive, but if it's a convenience store, then it's probably a building. It's better to spread the net wider with community OSM data. We can't rely on the building=yes tag being accurately added to everything
There was a problem hiding this comment.
Perhaps Rob knew something we didn't here - not sure why wheelchair, religion, place_of_worship etc is included for the education category. This one is probably good with your update
| "ways_line": [], | ||
| "relations": [] | ||
| }, | ||
| "keep": [ |
There was a problem hiding this comment.
What kind of output do you get in the tags when using this? I'm no expert in the osm-fieldwork / raw-data-api JSON syntax, but I thought the keep key was to specify what tags to keep on the final extracted features
| "facility:ventilator": [] | ||
| }, | ||
| { | ||
| "facility:icu": [] |
There was a problem hiding this comment.
things like this should definitely still be present: intensive care unit in a hospital, operating theaters etc
spwoodcock
left a comment
spwoodcock
left a comment
There was a problem hiding this comment.
Great start - thanks for getting this off the ground Anuj!
But I think it needs a bit more input into what tags are relevant, and probably some good testing (I haven't seen any evidence of tests, but I assume you might have done some - would be good to document 😄 )
As per our current workflow, we have this as a compulsory field. and if we only have osm based survey categories, it confuses users into thinking projects can be created for osm based surveys only. We can explicitly inform users that this also works for osm but generalizing can be more ux friendly instead of saying osm based categories only. |
|
I agree that the choice of tags will require more testing and additional input. |
What type of PR is this? (check all applicable)
Related Issue
Describe this PR
Changed enum names to remove
osmfrom survey categories.Updated the json config files for data extract such that the extracted data is more suited to the survey category chosen.