Venturing into…Social Media?

Kelvin Neo
Nov 7 · 5 min read

Finding Inspiration

After looking at land use, finance and Game of Thrones, I spent the past few weeks looking for another topic to dip my toes into. I wanted to pivot away from TV shows this time, and return back to something a bit more “real-world”. I thought of social media as a possible topic in a rather serendipitous way — a late-night conversation with a friend regarding our social media usage and presence on Instagram led me to think about what usage rates of different social media sites are like currently. For a generation so obsessed with social media, it is startling to realize that we don’t know much about the numbers behind it.

I first hunted around the internet for relevant data — I wanted a dataset that 1) spanned a long time so I can observe changes in key trends; 2) covered a wide spectrum of social media sites for a good sample size and 3) data that is updated till at least 2018. My initial thought that such data would be readily available given the ubiquity of the topic evaporated quickly — most of the latest and best datasets were locked behind paywalls. It took me quite a bit of time to finally find a Statista dataset that was updated and accurate!

Segment of the initial raw data

Cleaning the Data

Alas, this data set wasn’t flawless. I noticed that there was some cleaning to be done with the data: for example, there were a lot of missing years’ data for Reddit (2010–2013 missing), while Whatsapp’s 2018 data was missing. I had to manually search the web for this data — for Reddit, I went back to Reddit archives to check for their user numbers while I cross-referenced other sources for Whatsapp’s 2018 numbers.

Supplementing the gaps in the dataset

I eventually arrived at a cleaned version of the data — with which I could perform my initial analysis on! Using a PivotTable in Excel, I soon realized that with close to 20 platforms as data points, my graph was way too cluttered and messy. I decided to remove all the platforms that have already been phased out (e.g. Friendster) so the current data set only consists of existing social media platforms. Some quick superficial edits later, I ended up with this:

Preliminary graph of MAUs across social media platforms

There were 2 main things that I was surprised by initially: 1) mature platforms such as Facebook are still growing at such a fast rate?! 2) Reddit and Twitter having such a low number of MAUs.

Final Product

The legends were also not arranged in order and added clunkiness to the data viz — Prof. Burke gave the suggestion of removing the legend entirely and adding logos instead. I also adjusted the line colours to correspond with each logo colour using Format Painter for ease of reading:

Final Data Visualization with Prof’s recommendation of adding logos!

Prof. Burke also suggested looking at y-o-y growth trends: how have the changes in rate of growth changed over time?

Looking at y-o-y change instead

I thus expanded the dataset to look at change over time — unfortunately, the final product was way too messy and gave little insight. But playing with data is an iterative process after all…

Erm…

I decided to post the final product on Reddit to get some feedback.

Final product on Reddit!

159 upvotes, 45 comments. There was a healthy amount of discussion from users being surprised by the low numbers of users that platforms like Twitter and Reddit had, as well users who were astonished that poor PR by some platforms are still growing so rapidly.

Most importantly, there were users who gave me suggestions to improve my data visualization and thinking process behind my data — 1) removing chat platforms such as WhatsApp, because they do not qualify as “social media platforms”; 2) Adding some fast-growing platforms such as Tik Tok, Vine and Snapchat for a more comprehensive look.

I initially left the chat platforms in to highlight their meteoric ascent to mainstream usage — but did not realize how that leads to a less sharp thesis overall given the definition of social media. Definitely something important to improve on.

Conclusion

This was definitely a good first step into social media for me — and in the group project assigned by Prof. Burke, I also luckily landed up with Media (a whole new product there — check Wesley’s Medium article for our group’s reflection on that), giving me a chance to build upon this knowledge I got from this product. While I was sure hoping for this product to generate more discussion and learning points from Reddit users, the takeaways I’ve gotten as aforementioned are valuable. The key is to keep the data tied closely to my thesis.

Welcome to a place where words matter. On Medium, smart voices and original ideas take center stage - with no ads in sight. Watch
Follow all the topics you care about, and we’ll deliver the best stories for you to your homepage and inbox. Explore
Get unlimited access to the best stories on Medium — and support writers while you’re at it. Just $5/month. Upgrade