Forum - View topic
My data analysis project on the "Japanese Animation TV Ranking" series.




Anime News Network Forum Index -> General -> Anime
View previous topic :: View next topic  
Author Message
Shay Guy



Joined: 03 Jul 2009
Posts: 2653
PostPosted: Tue Mar 18, 2025 12:53 pm Reply with quote
I've been following this ANN column for over 15 years, and a few months back I got the idea to hone my data analysis skills by digging into the archived data. The project ended up kinda ballooning as I read more about Video Research and various stuff related to Python/Pandas/NumPy/etc., and there's still some improvements I could make -- improving my visuals, caching my DataFrames in addition to web pages (pd.read_html() is currently my biggest bottleneck for time), and so on -- but I'm fairly pleased with what I've got right now. You can check out my Jupyter Notebook here:

https://colab.research.google.com/github/ShayeHorwitz/anime_broadcast_ratings/blob/main/anime_broadcast_ratings.ipynb
Back to top
View user's profile Send private message
Tony K.
Subscriber
Moderator


Joined: 18 Nov 2003
Posts: 12080
Location: Frisco, TX
PostPosted: Tue Mar 18, 2025 1:01 pm Reply with quote
I applaud your effort. But I think very few people around here will know how to read this or what to look for. You might wanna' give a nutshell summary of what you were trying to do or what we should pay particular attention to.
Back to top
View user's profile Send private message My Anime My Manga
Shay Guy



Joined: 03 Jul 2009
Posts: 2653
PostPosted: Tue Mar 18, 2025 3:45 pm Reply with quote
Fair enough. The basics:

I collected, cleaned, and analyzed all the anime TV rankings published on ANN going back to 2007, and all the original Japanese versions going back to 1996. Where feasible (about 97% of rows), I also translated the titles used listed in the Japanese data. After that, I analyzed my dataset as best as I could to see if anything of interest came up. Among my findings:


  • Most of the top titles aren’t the names with wide recognition overseas. No surprise to those of us who follow those charts, but worth mentioning.
  • There used to be a lot of anime airing from 7-8 PM, especially on weekdays, that got big numbers. Most of those timeslots are gone now. Some big-name anime that aired in them moved to other timeslots over time — Detective Conan, One Piece, Doraemon, Crayon Shin-chan. Saturday morning and late afternoon ended up being where a lot of those ended up.
  • The diversity of titles in the weekly rankings has been highly variable, but it peaked in 2006 and has been on an overall downhill trend since.
  • Late-night anime is much more common on the charts than it was just a few years ago — back in the day, Nana was the only one that charted on a frequent basis. (Though the bulk of what we see there nowadays is Demon Slayer or Friday Anime Night, both of which air before midnight.)
  • Saturday morning is for preschoolers’ shows on NHK-E. Sunday morning is for action shows.
  • Live viewership numbers have dropped significantly over the years (no surprise).
  • Surprisingly, Detective Conan used to rank higher than Sazae-san several times a year back in the late '90s. In 2001 that dropped to twice, but it continued to happen once or twice a year through 2004. After that, the next time a regular anime episode (as opposed to a 100-minute-plus special) topped Sazae-san was in 2016.
  • The average ratio of “individual rating” to “household rating” has measurably increased since Video Research started reporting the former in 2020, possibly a mathematical result of declining household sizes.
  • Looking at average individual/household ratios for each year, there's some interesting patterns. They're consistent enough from one year to another that I’m convinced they’re meaningful, but I’m not sure how to explain all of them — I guess with Maruko-chan and the pre-Conan shows, there’s a sort of “opening act” effect where one person will be watching the first program, then another member of the household or two will join them once the more popular program after it starts?
    • Pokémon Horizons, Sazae-san, and the NHK-E kids’ shows have especially high ratios, which I guess reflects parents watching with their kids.
    • Soreike! Anpanman, Gundam: The Witch From Mercury, and shows in the pre-Conan slot have low ratios, as do most late-night anime (Demon Slayer’s is higher).
    • Detective Conan and Doraemon get similar ratios, with Doraemon’s a little higher most years; Pretty Cure gets higher ratios than them, but not as high as Chibi Maruko-chan’s, which are in turn lower than Sazae-san’s.
    • One Piece is fairly low by the standards of the family shows, but middle-of-the-pack overall.
  • The Digimon Adventure remake, Digimon Ghost Game, and Run for Money: The Great Mission have done really poorly by the standards of their timeslot.


I also ended up learning some interesting things along the way that didn't really make it in, like the history of the Nagoya TV/Sunrise programming block that moved a couple times over its 40-year history (1977 to 2017) and included everything from the Zambot 3, to the first three Gundam shows, to Aura Battler Dunbine and Heavy Metal L-Gaim, to the Yuusha series (Exkaiser to GaoGaiGar), to Dinosaur King, to Battle Spirits, to Tribe Cool Crew, before finally ending with Heybot -- ANN reported on it at the time, but it flew under my radar. It's a long and weird run. Or the timeslot Pretty Cure's used since it started, which it turns out has exclusively hosted Toei Animation shows since its first anime in 1984 -- the last non-shoujo series to air there was Ghost Sweeper Mikami.

My reading has also, in some ways, made me more confused about how Japanese networks work compared to American ones; I think I mentioned that in a Talkback post recently.
Back to top
View user's profile Send private message
Display posts from previous:   
Reply to topic    Anime News Network Forum Index -> General -> Anime All times are GMT - 5 Hours
Page 1 of 1

 


Powered by phpBB © 2001, 2005 phpBB Group