Vendor Recommendation

Inspiration

-- the vendors that fit you the best

What it does

-- Use Yelp's example data of business, user, and review to recommend vendors

How I built it

-- Data Preparation : Use Yelp's example data. It includes business data, user data, and user reviews -- Retrieval : Use pipecone to store the business category information, and use cosine similarity search to retrieval the related business that the user searches for. -- Ranking : train a recommender neural network using tensorflow recommender, and save the model, and rank the list of vendors from the retrieval stage

Challenges I ran into

-- The review data has 6 million rows, and I can only use 200k to train my model as google colab timed out

Accomplishments that I'm proud of

-- The system work!

What I learned

-- Various recommenders can be used to rank.

What's next for vendor recommendation

-- Build a UI

Built With

• pipecone
• tensorflow