- End to end example of Amazon Forecast using Snowflake.
- Note this example follows amazon-forecast-samples
- Uses simulated Snowflake as initial data store to for the source data
- After data transformation as needed by Amazon Forecast, data is updated back to Snowflake data store
- Unload data from Snowflake to S3 and import dataset to Amazon Forecast, build predictor and forecast
- Export Forecast Query and load it to Snowflake data store.
I recommend install Anaconda, a pre-packaged Python distribution that contains all of the necessary libraries and software for this project.
Once you have Anacoda installed, create a new conda env. In below command replace myenv with your preferred env name
conda create -n myenv python=3.6once you create myenv, activate it
conda activate myenv This use case requires Python 3.6 and the following Python libraries installed:
Code is provided in the Incorporating_Related_Time_Series_dataset_to_your_Predictor.ipynb notebook file. Some utility code is present in util folder.
Env file is present in env/dev/credentials file, this file shold contain snowflake connection and aws credentials as a dictionary.
In a terminal or command window, navigate to the directory advanced/Incorporating_Related_Time_Series_dataset_to_your_Predictor/ and run one of the following commands:
jupyter notebook Incorporating_Related_Time_Series_dataset_to_your_Predictor.ipynbThis will open the iPython Notebook software and use case file in your browser.