Context Based Grocery Recommendations

[DougEbel]

The data is from Kaggle: https://www.kaggle.com/competitions/instacart-market-basket-analysis/data

A video of the field person describing the video to Pranay: Gen AI Demo-20230522_103358-Meeting Recording.mp4

The notebook and python program are [attached:]

The Notebook is posted to files in our bi-weekly meeting: https://teradata-my.sharepoint.com/:u:/p/douglas_ebel/EevEUPA7mTNJnvFeDmS6RiEBo8LV-vXWPfUelkffkFIqHQ?e=Sf3Nsz

as is the Python program:

https://teradata-my.sharepoint.com/:u:/p/douglas_ebel/ETzvALFWsw5IoRyakeLYy24BL3AA3Vj5PkIpYDaF2pPgrw?e=sqWbDz

[DallasBowden]

Pranay want's this published, Alex S. has provided a new model but not providing adequate results.

[DallasBowden]

Pranay has commented and there are some edits that need to me made. A new table has been included and needs to be pushed to cloud. GSS1185 has a script that can be accessed from the monitor menu.

[ForbiddenDevil]

Created a pull request here

The previous branch was "grocery" but it was outdated and didn't have latest startup file. Hence created new branch "grocery_recommendation".

[DallasBowden]

Created a pull request using a new branch. There was an issue with the grocery branch. It was two behind the Main branch. Created a fresh branch. We need to delete the old branch "Grocery".

[DallasBowden]

Approach we are using is giving the recommendations. Only one or two is relevant. nPath is generating a different chart everytime. plot_firstmainparts is generating different results. We are sampling the top 1000. The random dataset will have different results.

[DallasBowden]

Updated nPath to always return the same rows. There is a new pull request.