TAHMO Incoming Solar Radiation Prediction Challenge
$10 000 USD
Under code review
Prediction
Geospatial Analysis
1525 joined
760 active
Start
Apr 01, 26
Enrolments close
May 24, 26
Close
May 24, 26
Reveal
May 24, 26
'/%3e%3cpath d='m450 224.315-44.467 136.87 116.401-84.559h-143.87l116.403 84.559z' style='fill:none;stroke:%23006233;stroke-width:14.62993431;stroke-miterlimit:4;stroke-dasharray:none'/%3e%3c/svg%3e){kind=small}
External datasets vs. train.csv only — how helpful are they?
13 May 2026, 21:23 · 3
Hi everyone, I’m curious about how much external datasets are actually helping in this competition. Has anyone managed to achieve a strong score using only the data provided in train.csv? Or are external sources proving essential for competitive results?
I havent tried external datasets yet, but with only the train csv , I have hit a ceiling at 0.31 plb score. I dont know what the 0.4xx and 0.5xx people are doing
did you achieve this with ensemble learning or with pretrained models ?
Indirect answer, but I made a post on the datasets I used, hope it helps.