victorlga/ML-ames-dataset

Exploratory Analysis of the Ames dataset.

Set up your environment

It is preferrable to use Anaconda (https://anaconda.org/) for this project.

Install Anaconda and then create the environment for this project with the command:

conda env create -f environment.yml

This will install an environment called "ames" with the latest Python for you.

Working in the project

Just activate the environment.

conda activate ames

Hugging Face Deploy

The "Ames" directory contains a hugging face deploy of an api that contains our Best model

To use it send a POST request to https://andrecorrea-ames.hf.space/predict with a json containing the processed model features in the request's body.

It's possible to test and try out the API through the link https://andrecorrea-ames.hf.space/docs

The documentation is available at https://andrecorrea-ames.hf.space/redoc

Self Host

Alternatively we can also self host the api by entering the "api" directory:

enter the api directory

cd api

install all packedges needed for the API

pip install -r requirements.txt

Then run the API

uvicorn main:app

The API will be running on http://127.0.0.1:8000/predict

For a easy way to test the API, you can go to http://127.0.0.1:8000/docs

The documentation is available on http://127.0.0.1:8000/redoc