Read on Twitter
Doing stylometric analysis on that Spokesman editorial at 2 AM because why not
Which, by the way, is another great reason why R is just better for data analysis than any other languages: it& #39;s the libraries stupid
Including https://github.com/computationalstylistics/stylo">https://github.com/computati...
Including https://github.com/computationalstylistics/stylo">https://github.com/computati...
Not expecting there to be enough data to make any hard statistical inferences, but the graphs can be fun
Using the methodology here [1], you can get closer to some claims of statistical significance though I should be noting I& #39;m using a black box for these computations and don& #39;t understand relations to sample size requirements and this is a 2AM shitpost
[1] https://computationalstylistics.github.io/blog/imposters/ ">https://computationalstylistics.github.io/blog/impo...
![Using the methodology here [1], you can get closer to some claims of statistical significance though I should be noting I& #39;m using a black box for these computations and don& #39;t understand relations to sample size requirements and this is a 2AM shitpost[1] https://computationalstylistics.github.io/blog/impo...](https://pbs.twimg.com/media/EWCanp_U0AAowpw.png "Using the methodology here [1], you can get closer to some claims of statistical significance though I should be noting I& #39;m using a black box for these computations and don& #39;t understand relations to sample size requirements and this is a 2AM shitpost[1] https://computationalstylistics.github.io/blog/impo...")
[1] https://computationalstylistics.github.io/blog/imposters/ ">https://computationalstylistics.github.io/blog/impo...
Also get an interesting result if you use the top 2000 words instead of the top 200
