An idea for a demo of a beer suggester app, based on smart-phone snapshots of beer taps in pubs
Making the world a better place, one beer recommendation at a time
- Google Vision API
- Google Image Search
- Google NLP API
- Pivotal Cloud Foundry
- Spring Cloud Dataflow
- (etc.)
A colleague was in London, walked into a pub, and was unsure of which beer to choose. Given this individual's past preferences, coupled with a photo of the beer taps in the pub, suggest the best choice.
Not sure. We need a demo in short order, so this could be too broad a task to tackle all at once. We may elect to narrow our focus somewhat.
-
Computer based image recognition finds the bounds of each of the beer taps, so they can be extracted and searched individually
-
Image based search returns the brand and type of beer, along with a link to a beer ratings web site
-
Rating data for this beer is scraped from the page and stored, keyed to an ID for this particular beer (form of ID should facilitate "fuzzy" search/retrieval)