Skip to content

Diverse recommendation tool for online beer shops that prevents customers from only receiving recommendations on high-sales beers, so that craft beers are not left alone!

Notifications You must be signed in to change notification settings

FeliciaMarlove/dont-let-craft-beers-alone

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

9 Commits
 
 

Repository files navigation

Project Title

Don't let Craft Beers alone!

Summary

A recommendation tool for online beershops and beer apps that doesn't rely only on sales volume and frequency. The idea is to prevent only recommending already well-known beer brands to allow customers to discover recent, local, craft beers.

Background

I'm Belgian and I love craft beer! My country has a rich beer diversity and a lot of craftmen and craftwomen making beer with passion. Still, craft beer appears to be something for the "beer geeks" while well established -industrial- bestsellers sell without problems. My fear on this matter is that most recommendation tools would give more weight to those beer giants because the sells level is higher, while beers should be recommended based on features like taste, variety... As Covid made the online shops, including food and beverage e-shops, more and more popular, recommendation tools can make a big difference. It should be as easy to discover craft beers in the aisles of a beer shop than being recommended a craft beer on such online shop.

How is it used?

This tool would typically be implemented on beverage webshops, beer webshops, beer recommendation applications or websites, social beer networks apps and websites, maybe on blogs about beer.

Data sources and AI methods

There are a few beer APIs available in order to collection data about beers, but data sets could be collection from future collaborators (webshops interesting in the solution or social beer apps).

Based on the example quoted in the Building AI course, "The Yle Areena content recommender", AI methods would be:

  • Collaborative filtering algorithms
  • Deep learning
  • Reinforcement learning

This has to be deeper thought about.

Challenges

  • While some data sets exist, I don't think they're complete.
  • I'm personally facing the challenge of being totally new to statistics, AI and Python (although I have experience in programming).

What next?

As a beginner, I would love to collaborate with one or two people motivated by this idea, with some more skills in AI and statistics. Python would be a great asset but that's the part I'm most confident in learning by myself (although I also aim to learn the other topics, I foresee my progression will be slower). I'd like to work with a small group or a peer, motivated to advance on the project on a pretty regular basis (better regular little steps than big erratic jumps). Maybe some day we'll raise a craft beer together to celebrate our working solution!

Acknowledgments

Building AI course project

About

Diverse recommendation tool for online beer shops that prevents customers from only receiving recommendations on high-sales beers, so that craft beers are not left alone!

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published