Hello Reddit! I made this map of the latest published research. Using the SPECTRE 2 embedding model and UMAP. Then a bunch of optimisations to try get it run smoothly. I was hoping it can help provide a macroscopic overview of research trends!
momoladebrouill on
This is so cool, it runs very smoothly on my phone browser, and it’s very interesting 👍👍
Wasteak on
This is awesome, its great to seet what topics are popular
cdig on
Wow it does a horrible job with volcanology research. Mostly in bio-agricultural sciences instead of physical science – related articles volcanology are pretty segregated.
thedrivingcat on
Hmm, having „Geochemistry and Geologic Mapping“ in the political science and law section seems like a coding mistake; the ‚top topics‘ are all political in nature and seem linked to democracy research.
surfergrrl6 on
Thank you for my new rabbit hole, OP.
Puzzled_Committee735 on
Really cool visualization. I work in the metascience space and this rocks. There are a bunch of journals that haven’t heard off (ever), so this may benefit from filtering a lot of low quality sources.
AnAccidentalRedditor on
There are many glitches in the matrix (some of them noted by other Redditors) but overall, simply wow! This is the epitome of beautiful data! 👏
hodyisy on
Why do so many things show as topic + AI?
Social sciences seem quite confused.
Also – great visualisation, looks amazing on Android phone.
Leave A Reply
Du musst angemeldet sein, um einen Kommentar abzugeben.
9 Kommentare
Hello Reddit! I made this map of the latest published research. Using the SPECTRE 2 embedding model and UMAP. Then a bunch of optimisations to try get it run smoothly. I was hoping it can help provide a macroscopic overview of research trends!
This is so cool, it runs very smoothly on my phone browser, and it’s very interesting 👍👍
This is awesome, its great to seet what topics are popular
Wow it does a horrible job with volcanology research. Mostly in bio-agricultural sciences instead of physical science – related articles volcanology are pretty segregated.
Hmm, having „Geochemistry and Geologic Mapping“ in the political science and law section seems like a coding mistake; the ‚top topics‘ are all political in nature and seem linked to democracy research.
Thank you for my new rabbit hole, OP.
Really cool visualization. I work in the metascience space and this rocks. There are a bunch of journals that haven’t heard off (ever), so this may benefit from filtering a lot of low quality sources.
There are many glitches in the matrix (some of them noted by other Redditors) but overall, simply wow! This is the epitome of beautiful data! 👏
Why do so many things show as topic + AI?
Social sciences seem quite confused.
Also – great visualisation, looks amazing on Android phone.