The second challenge of our dashboard week is to use YouTube trending dataset on Kaggle and create a dashboard in Tableau. This sounds more straightforward then yesterday. I thought I might be able to finish earlier today, but I was wrong. I had my biggest challenges when I was trying to get channel subscribers number as additional data.
I was trying to use YouTube API to get the supplement data I want. The dataset given contains trending videos data in 11 countries, which is quite large. As a result I used up all my API daily limit in one run. What’s worse is that I only realised this after I had to run my workflow again. After around 15 mins of wait I got a bunch of 403 Forbidden. My previous result was overwritten as well.
I’ve also tried to used another Google account to generate another API key. However it still didn’t work. And I guess that maybe my IP was already regarded as a robot. Then I’ve tried other ways to get subscribers data. Finally I decided to focus only on US market (considering the remaining time I have). And I get my supplement data on Social Blade. Data preparation almost took me whole day and it is something I didn’t expected in the morning.
My final workflows turned out to be quite simple. The first one is cleaning and parsing the given data. And the second one is more of data transformation for my supplement data to get it ready for Tableau.
Workflow 1: The given data
Workflow 2: My additional data
The story I’d like to tell is about whether the number of subscribers really that matter. I look at its relationship with other metrics such as number of trending videos and total number of views. Below is a screenshot of my dashboard. You can also view it on my Tableau Public.