Coke Studio & Data Analytics #3

In previous post, we concluded that last six seasons (post Rohail Hyatt) were major contributors in Coke Studio’s popularity. But we also concluded the things without involving any external factor in our analysis like Internet user growth in Pakistan, Coke Studio popularity on Facebook, advertisement expenditures by Coke Studio (Also highlighted by Kaneez Fatima ), Impact of other ventures like Nescafe basement, Pepsi battle of bands( Highlighted by Qais Khan ) etc. 

Let’s now focus on only one of the external factor: the internet users growth in Pakistan and how it helped other producers to bring on board bigger viewership than Rohail Hyatt.

We visualized the yearly internet users trend in previous post, which depicts that internet users jumped from 168K in 2008 to 74Million in 2019 i.e. 440 times growth in 11 years. Such a massive growth can't be ignored while making any conclusion about Coke Studio popularity as our analysis is purely based on Youtube statistics.

In below graphs we tried to visualize season wise video views and internet users in respective year and further two different dimensions in same graph.

We populated above graphs to find a correlation between internet users growth and Coke Studio's season wise views. But it dint led us to any correlation between both, so we are going to visualize again the same data but from different perspective. Now instead of year on year views, we will be visualizing accumulative sum of video views for each year. For example against year 2011, we ll add up the video views of all the seasons launched before 2011 and we ll try to correlate it with number of internet users in same year.

After visualizing the year wise accumulative Coke Studio views with yearly internet users, it became quite evident that Coke Studio viewership started to increase exponentially right after 3G/4G trials in Pakistan. Which is a strong evidence against our conclusion in previous post about Rohail Hyatt. As Coke Studio popularity after 2013 wasn't being driven by producers only but the significant popularity driving force was the internet users growth.

So we have to take a step back from previous conclusion and have to re-validate that “Rohail Hyatt contributed least in Coke Studio Popularity”. 

To re-validate our previous conclusion, let's now try to build a new KPI (Average Views per internet user). The new KPI can be easily calculated by dividing season views by the number of internet users in respective year. This derived feature is being calculated with an assumption that major portion of the video views were generated in same year i.e. For Season 3 significant chunk of viewership was generated in 2010, while trailing years didn't contributed much in viewership generation.

If we go through above populated statistics than we can easily say that Coke Studio during Rohail Hyatt had far better reach to internet users as compared to other producers. As the video views per internet user were highest during first 3 seasons of Coke Studio but this KPI latter started to degrade after Season 3. The latest insight clearly support my friends stance that Coke Studio content was the best only during the first few seasons, latter seasons couldn't provide much compelling content.

Now lets move forward by exploring the available data further in a quite interesting way. The new approach to data visualization will be based on simulating a scenario where we ll be considering constant number of internet users from 2008 till 2019 i.e. 74 Million throughout the decade. For simulation such scenario, we ll be populating a new derived factor for each year. This new factor will be simply created by dividing 74 Million with the actual internet users in each year. i.e. for 2014 the derived factor will be 14.3 (74.1M/5.1M).

Now in next step we ll be using this derived factor as multiplication factor to calculate a new viewership scale. PFB the updated/simulated viewership scale for each year, which was calculated by simply multiplying the derived factor with actual viewership of the season.

If we review above simulated statistics then we can easily conclude that Coke Studio popularity was at peak on internet, during the first three seasons and latter it started to decline gradually. Which again verify my friends assumption that best content of Coke Studio was only created during first few seasons of Coke Studio.

We can also use another quick test for our latest conclusion by comparing the latest songs of the Season 12 with any-other Pakistani content available on Youtube. For the subject test we will be comparing the Coke Studio song with one of the Pakistani Drama episode, both were released during the last weekend (16 Nov 2019).

Above two screenshots are enough to support our latest conclusion that Coke Studio has lost its popularity on internet and other content generators are currently more popular on Youtube. As a drama episode was uploaded 1 day ago but grabbed approx. 5 times higher views than Coke Studio Song, which was uploaded 3 days ago.

My next post will be all about outliers handling and how outliers can impact over all data analysis journey. Next post will be the last post in series & I am working in parallel on another thread, where we can gauge the viewership scale of Pakistani Drama Industry on Youtube.