I did a quick pass of Grab mapping team edits in Thailand. The intent is to understand whether there are egregious edits made by the team in the context of the recent TechCrunch article published last week. This is a quick evalutaion and does not capture the full picture. But it provides general trend. This evaluation is purely arm-chair and I did not do any ground truthing.
- Using OSMcha, we filtered changesets made by grab's data team in thailand - https://osmcha.mapbox.com/?aoi=739470d2-4cb5-4c40-93b8-205adc88f264.
- I inspected approximately 200 changesets.
Notes
Changeset too big - edits crossing several countries or too large to investigate. This type of editing should be avoided. Make small changesets as much as possible.
- 65553833, 65441984, 65364684, 65036912
Deletions by the same team - previously added roads (service) were deleted afterwards. Why did Grab team do this. In most cases the previous addtions were valid based on satellite imagery (either Bing, DigitalGlobe, ESRI and Mapbox).
- Examples - 65032561, 65031974, 65031327, 65018021, 65005873, 65004829, 65004986, 65004568, 65002609, 65001005, 64964666, 64963644, 64961749, 64960325, 64957490
JOSM validation - common editing errors found within the changeset bbox. No error was found most roads that were flagged were due to missing names.
- 64315343
- 64313144
- 64284700
- 64283020
- 64282915
- 64281932
- 64280172
- 64279419
- 64278626
- 64218076