Amini Soil Prediction Challenge
Helping Africa
$7 000 USD
Completed (~1 year ago)
Prediction
Earth Observation
1069 joined
336 active
Start
Apr 02, 25
Close
Jun 22, 25
Reveal
Jun 23, 25
'%3e%3cg id='b'%3e%3cpath id='a' d='M0-95V0h50' transform='rotate(18 0 -95)'/%3e%3cuse xlink:href='%23a' transform='scale(-1 1)'/%3e%3c/g%3e%3cuse xlink:href='%23b' transform='rotate(72)'/%3e%3cuse xlink:href='%23b' transform='rotate(144)'/%3e%3cuse xlink:href='%23b' transform='rotate(216)'/%3e%3cuse xlink:href='%23b' transform='rotate(288)'/%3e%3c/g%3e%3c/svg%3e){kind=small}
Discrepancies in PID Matching Between Sentinel-2 and Training Dataset
Data · 29 Apr 2025, 14:06 · 2
There are many missing PIDs in the Sentinel-2 DataFrame. When I try to find the closest point using longitude and latitude, the closest match for some missing points is too far away. Why is there such a large difference between the training dataset and the Sentinel-2 data? I need help understanding this issue.
There are lots of issues with the data in this contest. I don't think using lat/lon distance will be useful since the coordinates don't match the info given on the overview page. See this discussion; [location related query](https://zindi.africa/competitions/amini-soil-prediction-challenge/discussions/26202).
You can use the landsat data instead, as it is very similar to sentinel..it has enough matching IDs in the training and test set