Netflix TV shows watched in each country
?????????????? ??. ????????
Projects | Data Analysis | Product Development
A challenge: you′re presented with a dataset you have no idea about. You only know it′s related to the top 10 TV shows in Netflix per week and per country during from July 3rd, 2021 thru November 18th, 2023.
You download the dataset, clean it up a bit and you get this overview of what′s in front of you.
How do you go about to explore it? You don′t have any idea or objective yet. But remember: the data is there, whispering something to us. We just don′t know what it is about yet.
Tip 1: Ask questions: Which countries featured the more times in top ten weeks per country in the period?
In the dataset above, the info is daunting. Several columns, with country, country code, name of the show, Season, cumulative count of times in top 10 lists.
It begs the question: what country watched the most TV? Or which show featured in the top 10 the most per country?
Now you′re getting started...
Tip 2: based on tip 1, run a simple pivot table in Excel to see if your questions surface any pattern.
Now we′re starting to spot a pattern, rigth? Average times a show features in top 10 lists is ~5 times. And, depending on the country, a show is featured more than 100 times in top 10 (not show in the snapshot).
Bingo! I will go down this road.
领英推荐
Tip 3: repeat and rinse 1-2. When you find something interesting, put visualize the data.
If you are in a hurry, you can use MS Excel for this visualization, ending up with this chart:
So, here are the winning TV shows:
And how about United States? In the same period, the winner was...... CoComelon (never heard of).
Let′s try out another visualization, this time with PowerBI and its world map mode:
This visual is way better on the eyes, but require additional work to single out the TV shows. Note that the top 10 shows are highlighted (and corresponding country), with slightly different bubble sizes.
In the future I will dig more on PowerBI visualization tips.
Note: the possibilities to explore the data using the dataset are several. Other aspect and what I call "aha moments" can be derived from it.
Predictions could be made using forecasting methodologies with error margins.
One interesting example of prediction, extensively explored by the works of John Gott III, could be: given that a TV show featured 20 weeks on a specific country, for how long it will feature more?
====================================================