How does Netflix know what movies we like? Is it by asking us directly? Or maybe Netflix knows our best friends and asks them? I don’t think that’s how Netflix movie recommendations work. Today I will explain how Netflix provides a movie recommendation system using the correlation method.
Correlation
Correlation measures the degree to which two phenomena are related to one another. sounds a little complicated right? let me give an example. Imagine we are on a sunny beach when the temperature touches 40 C, it is very reasonable if many people feel hot with such conditions. At times like this people tend to cool their body temperature using cold drinks such as sprite, Coca-Cola, or maybe young coconut ice. It can be concluded that with the increase in air temperature will result in an increase in sales of cold drinks.
The graph presents the number of cold drinks sold on the y-axis and the temperature on the x-axis. It clearly shows that as the temperature rises, people tend to buy more cold drinks, leading to an increase in sales. This suggests a positive relationship between temperature and cold drink sales
Now imagine a student whose classmates with good grades are all taller than 180 cm. He firmly believes that a person’s height influences their intelligence — the taller the person, the higher the grades they achieve. Then he conducted research outside his class, resulting in a table as follows.
The graph shows the height of the student on the y-axis and the number of marks obtained on the x-axis. From this, we can see that we cannot conclude that one’s height affects one’s intelligence
For the last example, it is a negative correlation, meaning that one event affects the decrease of another event, for example, you own a car and over time the selling price of your car will decrease. This is due to the number of components that are damaged, and the decline in people’s interest in older car models.
The graph shows the increase in years on the x-axis and the value of the car over time on the y-axis. As the years go by, we can see that car prices tend to decrease.
So.. what’s the point?
After knowing correlation and its types. Now is the time for me to explain how the recommendations that Netflix has work. Basically, Netflix uses the concept of correlation. First, we rate the movies we have seen. Netflix then compares our rating results with other viewers who have similar correlation rating patterns to us. That way Netflix can develop a movie recommendation system that we like but haven’t seen.
If you enjoyed this post on Sentiment analysis and interpreting data through data viewpoints, feel free to get in touch with me (Febrian Nur Alam) regarding any thoughts or queries!