I don't subscribe to Spotify, so I don't get an end-of-year Spotify Wrapped summary. But not to be outdone, Last.fm offers a "Playback" report. It's basically the same thing as Wrapped except that Last.fm waits until the year is actually over before sending out the report.
I managed to scrobble a few more tracks than last year:
My Playback report isn't all that interesting so it's not worth sharing. What I do want to highlight from it is the need for people to learn when to use averages, and when to use medians. Here is one of the factoids my report tells me, that my top track was Right Back to It by Waxahatchee with MJ Lenderman. My 17 listens puts me in the top 5% of listeners of that song.
Next, the reports gives me this graph which shows the cumulative number of listens to the track by me and by the average of all users who also listened at least once. Despite being in the top 5% of all listeners, you'll notice that my line is below the average line. These two facts combine to illustrate why averages are often misleading, and medians are more informative.
Let's put it this way, if I was serving food with Warren Buffet at a shelter for 100 homeless people, on average we would all be billionaires. Then asking the question "on average you're all billionaires, why are most of you homeless?" abuses statistics. The median wealth in that shelter would be zero, which is far more accurate of the financial condition of most of the people there.
In the same way as a few people can be immensely wealthy and skew averages into meaninglessness, all it takes is a few people who put Right Back to It on repeat to heavily skew the plot into absurdity. It's likely that the median line for this track might be nearly flat at just a few plays. Maybe the people at Last.fm think that looks funny, but if I'm in the top 5% of listeners, my listening history should look very different from the median (most other) listeners!
Here's the end of year report I'm sending to Last.fm:
