I was invited as a speaker in Data Natives 2017 to give a talk on tech trends track and I chose to speak about Recommender Systems because it’s a topic that fascinates me. It’s a topic that’s changing the way we shop clothes, order food, purchase books and more!
Behind those systems, there is a big whole stuff being done by machine learning algorithms, analyzing your behavior from your clicks, views, likes, purchases, social network, localization!
One of those famous recommender systems is the one of Amazon.com. When you find sentences like “Recommended for you, Sarah” or “Frequently bought together” or “Customers who bought this item also bought..” you are actually coming across outcomes of these machine learning algorithms.
I know, it seems like…
But it’s not! and you can do it as well! you can build your own recommender system, you just have to get an answer for How to start? and what to consider when building it? and I came here to help you make your first steps in :).
I stated Amazon as an example because they are leveraging a method called “Collaborative Filtering” in their Recommendation Engine. This method is frequently used in recommender systems, it’s very famous to the point that sometimes people think that “recommendation engine = collaborative filtering” but Recommender Systems are not the only application of Collaborative Filtering, and recommendation systems can be done using other methods too.
Others like Netflix and Last.fm are using collaborative filtering too, among others..and thanks to that method.
“35% of Amazon.com’s revenue is generated by its recommendation engine” [source]
“More than 80% of the TV shows people watch on Netflix are discovered through the platform’s recommendation system.” [source]
What do you wait, let’s pull back the curtain on Recommender System using Collaborative Filtering!
Check my talk slides here.
And if you want to watch a video, here is the official video.
If you have any questions about the topic, don’t hesitate to post a comment and I will be happy to answer!