#MLibRes
Norms of science: communality vs. secrecy, universalism vs. particularism, disinterestedness vs. self-interestedness, & more #mlibres
February 4, 2025 at 4:52 PM
Trending tags used globally in the past 60 minutes:

1) #art
2) #photography
3) #nature
4) #birds
5) #wibooindication
6) #resist
7) #booksky
8) #furry
9) #mlibres
10) #furryart

#trendingtags #trends
February 4, 2025 at 5:01 PM
Thank you @cora_jr for an interesting @UMLibRes talk today! #mlibres
February 4, 2025 at 5:21 PM
Growing understanding that alternate, digital dissertations can be just as rigorous as the traditional dissertation #mlibres
February 4, 2025 at 5:21 PM
Beyond text analysis: visual analysis, audio analysis. Expand how we look at a topic or set of data. #mlibres
February 4, 2025 at 5:21 PM
Academics can be limited by the digitized texts that are available; accessibility of tools & methods #mlibres
February 4, 2025 at 5:21 PM
How can library help? Try to make databases & corpuses available in front of the right people. Repositories can hold prepub work #mlibres
February 4, 2025 at 5:20 PM
Hard balance to strike with openness and publishing. Easier if don't have to focus on traditional publishing methods #mlibres
February 4, 2025 at 5:20 PM
Haven't done much teaching of digital methods, but will have one-on-one convos with interested students. Takes time, trial & error #mlibres
February 4, 2025 at 5:20 PM
Put visualizations online is valuable, viral stuff is often simplest. Hard to put into "slick" product ppl can easily understand #mlibres
February 4, 2025 at 5:20 PM
Can b hard to talk to marginalized groups about this work. At least start the convo: Does this make sense to you? Collaboration #mlibres
February 4, 2025 at 5:20 PM
Call 2 action: lift up work by & for marginalized groups! (solicit! give ppl funding & space) open up data & process when possible! #mlibres
February 4, 2025 at 5:20 PM
Can be very hard to collect a corpus to do computational topic modeling on - must have the corpus and have digitized materials #mlibres
February 4, 2025 at 5:20 PM
why not more projects addressing cultural work by marginalized populations?? archiving challenges; lack of resources? skepticism? #mlibres
February 4, 2025 at 5:20 PM
Kim Gallon: black press research collective#mlibres

blackpressresearchcollective.org
Black Press Research Collective – Promoting research and Scholarship on Black periodicals
blackpressresearchcollective.org
February 4, 2025 at 5:20 PM
Aditi Muralidharan: slave narrative analysis; developed wordseer:#mlibres

wordseer.berkeley.edu
WordSeer Project Page – Project Page
WordSeer is a text analysis environment that combines visualization, information retrieval, sensemaking and natural language processing to make the contents of text navigable, accessible, and useful.
wordseer.berkeley.edu
February 4, 2025 at 5:20 PM
Tahir Hemphill did hip-hop wordcount#mlibres

tahirhemphill.com
Tahir Hemphill Studio
Tahir Hemphill Website
www.tahirhemphill.com
February 4, 2025 at 5:20 PM
Others have worked on topic modeling with slave narratives, black women's suffrage, black press #mlibres
February 4, 2025 at 5:20 PM
Also looked at regional distribution of topics, where they occur the most. Some interesting patterns emerged #mlibres
February 4, 2025 at 5:20 PM
a lot of people making claims about the use of "ghetto" vs. "hood" but topic modeling pulled out the actual trend #mlibres
February 4, 2025 at 5:20 PM
First set of hip hop topics: references to neighborhood & happenings; religion/christianity; childhood/family experiences #mlibres
February 4, 2025 at 5:20 PM
w/o any knowledge about the corpus, can pull out meaningful information even tho computer doesn't know relationships between words #mlibres
February 4, 2025 at 5:20 PM
Used method of topic modeling: creates topics based on the text-what words occur with one another #mlibres
February 4, 2025 at 5:19 PM