Otto Recommender System using Pyspark and ALS

This projects was done to practice pyspark and ALS concepts

This recommender system dataset belongs to german online retailer 'Otto' and was taken from the following kaggle competition

This datset has 12 million users, 2 million items and 200 million events (clicks,carts and orders)

Following tasks were performed

Converting data from original jsonl format to parquets
Training ALS model on the parquets. To save compute, rank and iteration is kept small at the cost of accuracy
Computing top 20 recommendations for items in test data

Name		Name	Last commit message	Last commit date
Latest commit History 7 Commits
Computing_recommendations_using_ALS.ipynb		Computing_recommendations_using_ALS.ipynb
Converting_chunks_of_jsonl_into_parquets.ipynb		Converting_chunks_of_jsonl_into_parquets.ipynb
Loading_data_Recomminding_most_popular.ipynb		Loading_data_Recomminding_most_popular.ipynb
README.md		README.md
Training_ALS_Model.ipynb		Training_ALS_Model.ipynb

Provide feedback