A tool to find good places for dating.
Create a product that helps people who are dating find a perfect place for their dating experience with interest-focused approach (criteria) instead of position-focused.
A product (currently thinking of an app or a website) that does not have cuisine as filter. Instead, it will focuses on leverage Google Place API to analyze user reviews using NLP and stats to develop interest-focused filters such as quietness, cleanness, service, intimacy, taste, lighting, price. Could also introducing feature to help couple select restaurant with data collection.
- Tested Google Place API using Python Client
- Build out basic client to fetch 10 reviews
- Implement client to fetch all reviews and ratings for butcher chef and another restaurant store into two json for testing kw model and to perform t-test
- Google Place API only return up to 5 reviews. Decided to leverage 3-rd party API called SerpApi
- Use SerpApi pagination token to get all the reviews
- Implement kw extraction class method using mock data
- Diversification using Maximal Marginal Relevance algorithm (maximizing the similarity with review embeddings while ensuring different from existing extracted keywords)
- Implement REST API endpoint and testing on Mock data
- Deploy using Docker, deploy in AWS
- quite count, loud count, review to rating ratio
- Format and document BE code and classes
- Start FE dev with mock json data
- Implement cloud storage act as cache