thinkingmachines / unicef-ai4d-poverty-mapping

UNICEF AI4D Relative Wealth Mapping Project - datasets, models, and scripts for building relative wealth estimation models across Southeast Asia (SEA)

https://thinkingmachines.github.io/unicef-ai4d-poverty-mapping

MIT License

20 stars 8 forks source link

Single country rollouts #162

Closed tm-jace-peralta closed 1 year ago

tm-jace-peralta commented 1 year ago

What does this PR contain?

Output for https://github.com/thinkingmachines/unicef-ai4d-poverty-mapping/issues/125, https://github.com/thinkingmachines/unicef-ai4d-poverty-mapping/issues/126, https://github.com/thinkingmachines/unicef-ai4d-poverty-mapping/issues/127, https://github.com/thinkingmachines/unicef-ai4d-poverty-mapping/issues/138

Adds folders for each country: Timor Leste (tl), Philippines (ph), Myanmar (mm), Cambodia (kh)
Adds notebooks for generating the populated grids 2_*_generate_grids.ipynb and rolling out the cross-country model over this grid 3_*_rollout_model.ipynb.
Adds pkl files of trained single country models <rollout_date>_*_single_country_model.pkl

The final output of these notebooks contain (1) the predicted relative wealth index and (2) binned wealth category from A-E

Notes on rollout runs

First time runtimes This can vary based on machine specs and time of run (due to server capacity of the source datasets)

Country	N grids	Grid generation	Feature generation
Timor Leste
Philippines	46483	48 mins	61 mins
Myanmar
Cambodia

Removed all of the output of the cells with .explore in all 3_*_rollout_model.ipynb notebooks to reduce overall repo memory consumption. An accompanying screenshot of what is expected after .explore is ran was provided instead.

review-notebook-app[bot] commented 1 year ago

Check out this pull request on

See visual diffs & provide feedback on Jupyter Notebooks.

Powered by ReviewNB