tsforecast 0.0.1

ML based forecasting

Forecasting

Install

First, clone the repo like and cd to its root path, like so:

git clone [e-mail]:JaumeAmoresDS/forecasting.git
cd forecasting

Then install dependencies using one of the following options:

Option 1: installing from forecasting.yml file

conda env create -n forecasting --file forecasting.yml
conda activate forecasting

Option 2: installing from requirements file

conda create -n forecasting python=3.10 pip
conda activate forecasting
pip install requirements.txt

Option 3: installing from setup.py

conda create -n forecasting python=3.10  pip
conda activate forecasting
pip install -e .[dev]

How to use

Run:

python scripts/run_pipeline.py

This will save the predictions into a file called predictions.parquet. To display it, you can do something like:

import pandas as pd

predictions = pd.read_parquet ('predictions.parquet')
predictions.plot ();