# Extension Suggestion: Corpora Rankings

**URL:** <https://community.wanikani.com/t/extension-suggestion-corpora-rankings/16270>\
**Category:** API And Third-Party Apps\
**Created:** [March 5, 2017, 12:29am UTC](https://community.wanikani.com/t/extension-suggestion-corpora-rankings/16270 "2017-03-05T00:29:15Z")\
**Posts on this page:** 4\
**Page:** 9

<div class="post-metadata">

**Author:** ![StellaTerra](https://sea1.discourse-cdn.com/wanikanicommunity/user_avatar/community.wanikani.com/stellaterra/32/7851_2.png) [@StellaTerra](https://community.wanikani.com/u/StellaTerra)\
**Post date:** [March 26, 2017, 8:09pm UTC](https://community.wanikani.com/t/extension-suggestion-corpora-rankings/16270/161 "2017-03-26T20:09:47Z")

</div>

Not dead!

Life is going to be extremely complicated through April, so I wouldn’t expect any more progress for a bit after this, but I have done some work this morning that I wanted to share with y’all.

I combined all the vocab data that I want to use (in what I guess will be my version of this extension) and I’ve got it to a permanent public host. In case anyone wants at it, it’s [here](https://s3-us-west-2.amazonaws.com/wkcorpora/wkcorporadata.json). I’ll be updating it in that location as I move forward, but it shouldn’t be any problem to take on the bandwidth costs from that, as it’s S3, so it’s very cheap!

Also, I’ve been playing with the data. I had this curiosity about what the data would show if I took one of the corpora and built a graph demonstrating the distribution of the ratings of all of the words in a particular WK level, and then stringing them all together. So here’s what it looks like:

 ![](https://global.discourse-cdn.com/wanikanicommunity/original/3X/2/8/284b6dd4d82b3403ee23dd53a620593f652a41aa.png)

It’s a bit hard to read, but you can generally see that the higher rating words become fewer and fewer as you go up in levels. It seems like there is a real turning point at around level 20-25 where it turns from being mostly 3 star and greater to being mostly less.

Let me know if you find this stuff interesting and I can post other corpora’s data.

❤

---

<div class="post-metadata">

**Author:** ![CaoxKitty](https://sea1.discourse-cdn.com/wanikanicommunity/user_avatar/community.wanikani.com/caoxkitty/32/5660_2.png) [@CaoxKitty](https://community.wanikani.com/u/CaoxKitty)\
**Post date:** [March 30, 2017, 6:33pm UTC](https://community.wanikani.com/t/extension-suggestion-corpora-rankings/16270/162 "2017-03-30T18:33:01Z")

</div>

Good work guys! がんばって

---

<div class="post-metadata">

**Author:** ![1ember](https://sea1.discourse-cdn.com/wanikanicommunity/user_avatar/community.wanikani.com/1ember/32/6036_2.png) [@1ember](https://community.wanikani.com/u/1ember)\
**Post date:** [March 30, 2017, 8:56pm UTC](https://community.wanikani.com/t/extension-suggestion-corpora-rankings/16270/163 "2017-03-30T20:56:38Z")

</div>

I had some inspiration on how to approach the issue of the Anime corpus. You guys may have already come up with this, but I offer it here for your consideration:

Subtitle files

Such files tend to be available for free from various, if admittedly sketchy, 3rd parties, and could be used to build a solid frequency list with minimal parsing.

Really looking forward to the finished product!

---

<div class="post-metadata">

**Author:** ![StellaTerra](https://sea1.discourse-cdn.com/wanikanicommunity/user_avatar/community.wanikani.com/stellaterra/32/7851_2.png) [@StellaTerra](https://community.wanikani.com/u/StellaTerra)\
**Post date:** [March 31, 2017, 12:24am UTC](https://community.wanikani.com/t/extension-suggestion-corpora-rankings/16270/164 "2017-03-31T00:24:13Z")

</div>

Yup. That was the plan. If you look at the original post I wrote “Anime/Drama subs”.

I only know of the one site… [kitunekko.net](http://kitunekko.net), but if you have more suggestions about where to get reliably good subs I would be very grateful!

[Previous page](https://community.wanikani.com/t/extension-suggestion-corpora-rankings/16270.md?page=8)