Closed Niklewa closed 8 months ago
I have pulled all the most recent changes from the target branch. I have run the cleaning pipeline for all the files because there were some counties lost along the way (just a couple of them). I have cleaned the health
variable, but ultimately, I have not introduced it since it would result in losing around 15 counties; I'm not sure if it is worth it.
Some tests are failing, and I have checked them. I believe they fail because the freshest version of our dataset is not compatible with the trained causal framework. They became an issue after running the entire pipeline.
test_transformed_intervention_from_percentile_accuracy
: It fails as it's comparing the stored output of the intervention with the most recent one.
For very similar reasons, the last 3 tests in test_inference
fail, and along with them, causal_insights_demo
fails too.
In this PR, I have worked on variables from the areas of climate hazards, housing and energy burdens, health, and age distributions.
The major issue is that in the upcoming PRs, which have not yet been merged, there are changes to the files containing functions intended for cleaning in this PR. Firstly, it needs to be pulled from branches that are to be merged, and then the cleaning pipeline should function correctly (I do not want to modify those functions in this version to avoid conflicts).