Your Recommendations

You are almost there, so don’t give up! Before you start coding your final part for the content-based recommendation, have a look at how the output could look like. So in our example, Tobis last watched movie/show was Breaking Bad, and as recommendations, we got:

There are some very interesting recommendations, but most of them make sense, so it worked! Now it’s your turn!

One of the challenges is that not every public data set might include your last-watched movie/show. That is why we worked with a simple for loop, which iterates over your last seen movies/shows and first checks if the public data set includes that movie/show, and if so, you can build your data frame as mentioned in theory:

So you take out of each similarity matrix (description/genre/actors) the column of your last seen movie and create a final similarity column. As already mentioned, you can weigh the aspects (description/genre/actors) differently and check what happens to your recommendations.

For example, if you say that description, genre and actors are all equally important to you, then you can choose for each a 1/3 share:

For instance: I have looked at recommendations for “Inception,” and one of them was “Lord of the Rings,” but once I changed my priorities/weighting, that recommendation changed.

If the public data set doesn’t include your last-watched movie/show, your loop should move on to the next movie/show you have watched.

Since you only want to print recommendations for 5 movies/shows, you also need to define that your loop stops after movie/show number 5. Don’t forget to sort your data frame based on the final similarity column since you want recommendations with the highest similarity.

Voilà! You got yourself a self-built content-based recommendation system!

Congratulations! You’ve made it to the end of your TechAcademy Data Science project. After visualizing the data in the first part, you’ve also set up your personal content-based recommendation system in this section. If you stuck around until this part and managed to code the better part of the exercises, you have definitely earned your certificate! We hope you had fun learning Data Science with this combination of data sets and enjoyed it – at least the parts where you did not get stuck forever because of some unexplainable coding error. Do not forget to send your project results to our project submission email address before the deadline (04.02.2024, 23:59). Thank you for being a part of TechAcademy!