AAG-project’s Documentation

Usage:

Add the initial data to a “data” folder inside the pipeline folder. This is the initial data provided by AAG that has to be cleaned. Using the provided conda environment, run the pipeline to create a final parquet file that is used for the next part, clustering.

Contents:

• Main entry point
  • customerPrediction()
  • itemPrediction()
  • main()
  • preprocessData()
• Preprocessing
  • addFeatures()
  • blendPostalCodes()
  • finalCleaning()
  • initialCleaning()
  • mergeOnKunden()
• Customer clustering
  • clusterRFM()
  • getTop25PercentCustomers()
• Customer predictions
  • predictRevenuePerCluster()
• Item clustering
  • clusterRFC()
  • getDistributionCentres()
  • getItemDataset()
• Item predictions
  • predict_quantity_per_cluster()
• Database preparation
  • process_lstm_files()
  • read_and_process_excel()
  • save_to_sqlite()
• Helpers
  • convertDataToParquet()
  • fillMissingStates()
  • findNearestState()
  • getClosestMatch()
  • loadInitialData()
  • loadParquetFile()

Indices and tables

• Index

• Module Index

• Search Page