ZacharyST
@ZacharyST
Followers
817
Following
1K
Statuses
2K
Prof. of public policy at UCLA's Luskin School of Public Affairs. Protests, social networks, data science. Enjoy fresh coffee, the NBA, and city walks.
Los Angeles, CA
Joined November 2008
Rejected from NSF MMM with good/fair, good, very good, good, very good/good, excellent, excellent, and fair ratings. How do I interpret that?
0
0
0
RT @AmandaMurdie: Oh, boy, friends, I've got GREAT NEWS to share. Dr. @IdeanSalehyan - likely the world's leading peace science scholar stu…
0
3
0
RT @polanalysis: Now in FirstView: “Improving Computer Vision Interpretability: Transparent Two-Level Classification for Complex Scenes.” @…
0
7
0
Happy to see this formatted for print, a nice offset to several rejections last week. Don't forget to check out the API!
#OpenAccess from @polanalysis - Improving Computer Vision Interpretability: Transparent Two-Level Classification for Complex Scenes - - @stefan_scholz_, @nils_weidmann, @ZacharyST, @eda_keremoglu & Bastian Goldlücke #FirstView
0
0
1
RT @CUP_PoliSci: #OpenAccess from @polanalysis - Improving Computer Vision Interpretability: Transparent Two-Level Classification for Comp…
0
3
0
My favorite part of the holiday season is that, with the holiday break looming, all the rejections come in at once.
0
0
2
RT @davidhogg111: It’s really not complicated. The answer to the housing crisis is to build housing.
0
285
0
Is there an easy source of data on #nba players points per game or total per season? I want to teach regression using LeBron's trends in career points. is ok but I'd prefer not to pay, surely someone else has done this.
1
0
0
RT @matmagnani: #toread "Recommendations for sharing network data and materials", by @zpneal,@zack_almquist,@bagrow,@aaronclauset,@janadie…
0
4
0
👇👇Current CV transparency rests on interpreting pixels, not v. useful. We show how to use actual objects (hats, people, etc.) to understand CV classification. Read the whole thread, check out the paper, test the demo, and use the API.
New paper forthcoming in @polanalysis, w/ @nils_weidmann, @ZacharyST, @eda_keremoglu and Bastian Goldlücke! We propose a method that makes image classification more transparent by identifying which objects on the image are related to the outcome. Preprint:
0
0
2
Wow, super cool dataset: New dataset of daily global geolocated sentiment from 2019-2023 at admin-2 level for 164 countries.
0
4
27
RT @fgenovese__: Increasingly allergic to the quiet polisci norm that empirical papers should come up with elaborated theories/expectations…
0
52
0
This research could not happen now. 1, the streaming API, how we obtained the orig. tweets, is effectively dead. 2, we got users' tweet history thx 2 the Academic Research product, which gave us users' geo and non-geo tweets. Musk killed that product.
0
0
6