Surprise! I have new data, but it works great!

“Harvard Law Review (HLR) is the top ranked law review according to the 2020 Washington & Lee Law Journal Rankings. Law reviews are the preferred venue in which the majority of legal scholarship by law professors is published.

Research Questions:

Can a LDA model provide interesting insights into a wide variety of law review articles published by HLR?

How have Harvard Law Review articles changed over time? (Since 2000)

My data consists of every law review article that Harvard Law Review has published since 2000. This consists of 214 articles. I turned the data into a corpus and then cleaned it by removing stopwords, making the words lowercase, and removing numbers. This resulted in a cleaned corpus with 4,881,602 tokens.

After getting my topics I interpreted and labeled them as such : Topic 1: Sex/Gender Topic 2: Voting/Contracts Topic 3: Technology Topic 4: Insurance(Originalism?) Topic 5: International/Financial Topic 6: Regulation Topic: 7: Prison System Topic 8: Conflicts/Wars Topic 9: Classified Topic 10: Property/Ownership

”

Justin Burnworth
2022-04-26 
INFO  [23:36:56.049] early stopping at 100 iteration 
INFO  [23:36:57.899] early stopping at 50 iteration 
# A tibble: 10 × 10
   `Topic 1`   `Topic 2`     `Topic 3`   `Topic 4` `Topic 5` `Topic 6`
   <chr>       <chr>         <chr>       <chr>     <chr>     <chr>    
 1 transgender tying         facebook    harmless  copyright shop     
 2 contempt    subsection    software    absurdity commander epa      
 3 pregnancy   redistricting cyberspace  insurers  sharehol… dell     
 4 couples     taint         platform    linguist… sharehol… apa      
 5 biological  dilution      platforms   venture   cil       interage…
 6 parenthood  retributive   boilerplate valuation music     oira     
 7 nonbinary   shaw          balkin      elastici… slippery  fda      
 8 tribal      spatial       fiduciary   dictiona… ats       doj      
 9 mothers     districting   pc          original… ietf      omb      
10 marital     buyers        microsoft   insurer   sosa      sharehol…
# … with 4 more variables: `Topic 7` <chr>, `Topic 8` <chr>,
#   `Topic 9` <chr>, `Topic 10` <chr>