Research Ticket: User/Wine Recommendation Algorithms

[KatherineEvans]

Research some potential statistics and comparison metrics you'd be interested in tracking for your users:

Would a user be interested in seeing:

• Total number of wines they've posted about
• Total wines liked / disliked
• A profile of the average "liked" wine, a profile of the average "disliked"? (Is this a feature you want?)
• Percentage of posted about wines that are vegan/organic/sparkling/blends/types, etc.
• A method/algorithm to find users with similar stats to recommend - AND wines(?)

[michaelrmartin]

Ruby gem for recommendations - https://github.com/ankane/disco

Mosst promisinng

[michaelrmartin]

Most complicated ? - https://towardsdatascience.com/a-complete-guide-to-recommender-system-tutorial-with-sklearn-surprise-keras-recommender-5e52e8ceace1

[michaelrmartin]

https://medium.com/@ysmiracle/build-a-recommendation-engine-in-ruby-f6c4b44f2916

[michaelrmartin]

Bottld - uses python and nearest neighbor algorithm. might be a nice v3 option

https://github.com/FSA-Wine/bottld

[michaelrmartin]

https://www.geoffreylitt.com/2017/03/19/building-a-rails-recommendation-engine.html