The purpose of this tool is to aid in obtaining quick visualizations for time series forecasts provided by a Lightwood predictor.
At the moment, the tool supports predictors trained with lightwood >= 1.0
, but support for 3rd party models is coming soon~ish.
For now, there is no documentation as the package itself is fairly minimal.
However, most functionality is showcased through examples. Please refer to:
example/train.py
to train a Lightwood forecaster for airplane arrival data (which includes 4 different time series)example/visualize.py
to plot predictions from this model in your web browserexample/visualize.ipynb
to plot predictions from this model inside a jupyter notebook
Note: if you've cloned the repository (as opposed to pip install
ing), make sure the path to mindsdb_forecast_visualizer
is added to your python path environment variable before running these scripts from the package root folder.