Does everybody Listen to Christmas Songs only in Winter?
By Soumyadip Pal
It’s November already; 2016 is about to draw to a close! While in India, our major festivities for the year are over with Diwali, in the west, the holiday season is only beginning – there’s Thanksgiving this month, and then the big one, Christmas, to close the year.
There’s a nip in the air already – you can feel the winter coming. This is my favourite time of the year. The wintry chill last evening, reminded me of Christmas carols. Playing audio CDs, and even compressed music files from external devices are so passé, music today is all about streaming. I hooked my computer up with a wireless speaker, and started playing some lovely carols off YouTube.
Just then, I thought, surely I couldn’t be the only one with this November nostalgia! The good thing (or bad thing, depending on your perspective) about streaming media is that it has made a bouquet of data points easily available and accessible to anyone. I could, for example, check on YouTube, if people listened to Christmas carols all year round, or just during winter. My bets, of course, were on just the winters. Surely I couldn’t be that much of an outlier.
The first one that I listened to was titled “Christmas Songs for Children” from “The Christmas Channel”. I was curious to see the play statistics of the video, and as expected, there was a very strong seasonality with a massive peak around the end of December and early January. People started to play this video around December, with a peak at Christmas, followed by an abrupt fall as people went back to work. The Christmas Songs would then be forgotten, until the next Christmas.
An interesting side-note: the 45 minutes and 10 seconds track compilation had been played for a cumulative total of 98 human years, since it was published in October, 2013!
It’s quite understandable for Christmas carols. After all, in the peak of summer, who in their right mind would be in a mood to listen to them? I wondered if pop songs with Christmas themes too exhibited a similar seasonality.
The first song that came to mind was “Last Christmas” by “Wham!”. The statistics were, once again, on very expected lines.
I checked the statistics for two more similarly themed tracks, and the patterns were nearly identical.
While the total play count for the carols were more-or-less identical over two years, the play counts of the three pop songs kept increasing year-on-year. In other words, for each year, many more people watched the videos during December compared to July, and the December views too kept growing every year. Many more people watched the video last year, than the year before. This, perhaps, occurred simply due to the phenomenal growth of YouTube users, driven by the growth of Android.
The patterns for these views are what we call a time-series data. Such data measures a particular variable (in the case of these videos, views) over equal measures of time on the horizontal axis.
The presence of predictable, repeating patterns that recur annually is called seasonality in the data. In the case of the Christmas pop videos, there is a base effect, and the effect of the seasonality. The base effect is a continuously rising, however, the seasonality can either depress the base down to zero, or can boost it to very high number. The year-on-year growth in December is due to the increase in the base effect, or an increasing trend. More smartphone users, more people consuming data, more people consuming streaming videos in general. These, coupled with the seasonal effect of Christmas, creates the patterns that we saw above.
Of course, there would always be that one rebel who would watch a Christmas video in the peak of summer. Time-series even has that one covered. Statistics calls that “random effects”.
>> To get detailed information on Business Analytics with R, or if you are looking for a certification in Business Analytics using Hadoop please go to Manipal ProLearn.
Soumyadip Pal is a retail analytics professional and a passionate educator with more than 8 years in the industry and more than 7 years in the academia, currently working as a consultant with Manipal Prolearn.