Plex/Tautulli What's Your TV/Movie Viewing Age

[mcclown]

This is a much bigger feature than the other pull requests I've added.

This year's Spotify Wrapped added a feature to calculate your listening age....I've implemented something similar for Plex/Tautulli. It also includes a graph showing your viewing, by year, over the previous year. TVs and Movies are processed separately, so if you have enough data for both, you will see two age estimates.

Movie Card

TV Card

A few details about the implementation:

• Movies are calculated using the release year of each movie watched
• TV Shows are calculated using the release year of each episode watched (not just the show's premiere date)
• The calculation includes a logarithmic weighting of content to manage the bias for more recent content to be watched more often.
  • Older content has more weight.
  • We also include in the calculation how often a movie/episode is rewatched
• The visualisation shows the percentage of content that has been watched, per year. The chart is also scaled logarithmically to make it easier to read when individual years account for the vast majority of viewing.
• There is no support for music, yet. I've written the code to be as generic as possible, so it should be relatively easy to add that later. I have no test data for music because I don't use Plex for it.

[mcclown]

I forgot to mention the methodology for working out the year, once we've identified the median year from the weighted analysis of the viewing history.

The general theory seems to be that people's tastes are established between the ages of 10-25 years of age. So we use 18 years old as the rough middle of that. When we calculate the peak nostalgia year we then take assume that the user would have been 18 years old during that year...and take away 18 years to get their estimated birth year.

It's not amazingly scientific but it seems to be getting approximately right answers for most of the users I've tested with....or also hilariously wrong answers for some others. A conversation starter either way.

[aunefyren]

Wow, a ton of great work here, well done. Based on my chart I was born in 1980's. Is the logic because I watched a lot of movies from the 2000's in the past year, and if I was 18 in the 2000's, I would have been born in the 80's?

I was however born at the very end of the 90's, and my nostalgia movies are from the 90's and 00's. Not sure if that applies to everyone, but my nostalgia is not from when I was 16-20 years old, but rather from when I was 6-14. We need to generalize as much as possible of course, and I am known in my circle for my odd movie interests.

[mcclown]

Yeah, so I'm using the age of 18 as the year of peak nostalgia (as a proxy for people's tastes usually developing from the ages of 10-25, and 18 is right in the middle). Also, the movies/TV shows are weighted using a logarithmic weight, depending on their age...so basically an older movie has more impact than a newer movie. I did this to offset the fact that a lot of people tend to watch new releases of TV or movies. For TV shows, I based it on what episodes you watched the most, and the year they were released...so if you really like Season 1 of a show, it will take that year into account.

Not very scientific, it definitely won't be accurate for everyone, but my friends enjoyed it as a fun feature in last year's wrap.

...and a lot more accurate than any results I saw for anyone in Spotify. Spotify says I'm 82 but my movie age above is almost right.

[mcclown]

Just to call out, I haven't implemented it for music, as I don't host a music library. I tried to code it so most of the placeholders to support music are there but if you have the data you may want to take a look at finishing off music support.