There are NaN in the dataset we tried but failed to fill all of them.

• Create S3 bucket or Drive folder
• write code that uploads (water.model, light.model, heat.mode) to S3 or Drive. (If something already exists this replaces that)
• create a get endpoint for each environment. when this endpoint is hit the following things happen.
• already existing model in tfmodel directory gets deleted. new model (of the same name) is downloaded from S3 bucked or Drive folder and put into tfmodel directory. at the start of application

• API + Plotting
• Frontend Plot Tableau

• Get it deployment ready. (Docker, Heroku, Etc)
• API Setup
• Pydantic Schema and Example
• 15 get endpoints for actions. Initially only make 3
• 15 get endpoints for environments. Initially only make 3
• Placeholder model with 39 or less parameters