Neal Grantham 🆖 (@nsgrantham) header
Neal Grantham 🆖 (@nsgrantham) profile picture

Neal Grantham 🆖

@nsgrantham

In Bayes We Trust.

 San Francisco, CA
nsgrantham.com
Joined June 2013
  • Tweets1.9K
  • Following797
  • Followers1.2K
99Media
Photo by Neal Grantham 🆖 (@nsgrantham)Photo by Neal Grantham 🆖 (@nsgrantham)Photo by Neal Grantham 🆖 (@nsgrantham)Photo by Neal Grantham 🆖 (@nsgrantham)Photo by Neal Grantham 🆖 (@nsgrantham)Photo by Neal Grantham 🆖 (@nsgrantham)Photo by Neal Grantham 🆖 (@nsgrantham)Photo by Neal Grantham 🆖 (@nsgrantham)Photo by Neal Grantham 🆖 (@nsgrantham)Photo by Neal Grantham 🆖 (@nsgrantham)
extremely irresponsible to release a tool like this without providing any information on its false positive rate how many students will be falsely accused of plagiarism because their educator used GPTZeroX and believed it would "Detect AI Plagiarism. Accurately" as it claims?
Edward Tian (@edward_the6) profile pictureEdward Tian
@edward_the630 Jan 2023
We just launched GPTZeroX, a new AI detection model built for educators after a month of iteration. Demo👇
7
35
173
28,494
this isn't a benign application of machine learning , like detecting whether an image contains a cat or not — a wrong prediction means accusing an innocent student of cheating with the illusion of objective evidence ("the plagiarism detector said so"). What is the FPR?
1
29
1,049
tired: try guys wired: tri guise (the 2-simplex in 3-dimensional space)
A geometric figure of a 2-dimensional triangle embedded in 3-dimensional space with vertices at (1, 0, 0), (0, 1, 0), and (0, 0, 1).

ALT A geometric figure of a 2-dimensional triangle embedded in 3-dimensional space with vertices at (1, 0, 0), (0, 1, 0), and (0, 0, 1).

1
2
mbtaviz.github.io is an outstanding collection of interactive public transit data visualizations spatial maps, heatmaps, line charts, bar charts, horizon charts, and scatterplots — the gang's all here, each with only a slight touch of aggregation great work @msb5014 👏
Two visualizations of public transit data in Boston. On the left is a subway map with each individual vehicle marked by a dot and colored red, blue, or yellow depending on which transit line a vehicle belongs to. On the right shows time ascending downward on the y-axis and station on the x-axis. This leads to diagonally sloping lines down and to the right for vehicles moving in one direction, and diagonally sloping lines down and to the left for vehicles moving in the opposite direction. In total, the two visualizations communicate how frequently vehicles run in each transit line and how quickly they move from one station to the next, and where they may be unexpected slow downs and delays throughout the day.

ALT Two visualizations of public transit data in Boston. On the left is a subway map with each individual vehicle marked by a dot and colored red, blue, or yellow depending on which transit line a vehicle belongs to. On the right shows time ascending downward on the y-axis and station on the x-axis. This leads to diagonally sloping lines down and to the right for vehicles moving in one direction, and diagonally sloping lines down and to the left for vehicles moving in the opposite direction. In total, the two visualizations communicate how frequently vehicles run in each transit line and how quickly they move from one station to the next, and where they may be unexpected slow downs and delays throughout the day.

Jedwin Mok (@jedwinmok) profile pictureJedwin Mok@jedwinmok4 Sep 2022
Most useful thing ever, spend hours on this understanding space-time graphs and headway management mbtaviz.github.io/
Visualizing MBTA Data

Visualizing MBTA Data

An interactive exploration of ridership, congestion, and delay on Boston's subway system.

mbtaviz.github.io
1
25
opened a linear algebra textbook from my undergrad days and surprised how few pictures there are, beyond the obligatory "vector addition parallelogram" why isn't visualization more widely used in pure math textbooks?
1
1
4
my 3 best guesses are: - a preference for presenting the material in its most general form, "untainted" by concrete examples or contrived applications - insufficient software, especially for 3D graphs - "writing is hard enough, you're telling me i have to make pictures too?"
1
R users, what is your favorite package that interacts with a web API? I’m making an R package for the Discogs API and have so far relied on spotifyr as a source of design inspiration, could use a few more high-quality examples like it to draw from #rstats
7
12
64
when someone tells you that data cleaning has a "job relevance" of "4.0% of all DS jobs" that's a dead giveaway they have no clue what they're talking about aspiring data scientists: ignore garbage like this
Avi Kumar Talaviya (@avikumart_) profile pictureAvi Kumar Talaviya@avikumart_10 Aug 2022
TOP 10 Data science techniques you must learn to succeed in your data science career🔥📊🚀 A 🧵↓
12
tell me you don't use the terminal without telling me you don't use the terminal
This quoted post is unavailable.
Try viewing anyway · View on X
10
think you know Bayes’ theorem? stumbled upon this formulation in Schervish’s Theory of Statistics, and, uhhh
Image from Neal Grantham 🆖 (@nsgrantham)
1
1
4
always knew I should have paid closer attention in measure theory (cherry on top: this is page 16 of 600+)
be kind to your global environment, please do not do this
J. Mouchnino (@JulienMouchnino) profile pictureJ. Mouchnino@JulienMouchnino16 Jul 2022
Did you know that you can pull out a variable from a pipe workflow in R without interrrupting the stream of pipes ?💡 #dplyr #RStats
Image from J. Mouchnino (@JulienMouchnino)
2
10
data science advice that i continue to have to relearn: assume any plot you make will be shared far beyond the email / report / slide deck in which it originates does the plot include all the appropriate context for a newcomer to interpret its message correctly?
1
1
1
38
live look at my data pipeline
This Post is from an account that no longer exists. Learn more
5
"today, life won" — bullshit, thousands of pregnant women will die preventable deaths because they no longer have access to safe and legal abortions
1
6
and this comes a day after striking down gun control legislation in states like NY and CA, one month after the Uvalde shooting? no, "life" is not winning, not even close
2
so many medium posts on towards data science end before they actually get interesting you gotta do more than copy and paste the "getting started" section of a project readme and add a terrible stock photo on top, cmon
4
22
no better way to spend a saturday morning than fighting for my life in python environment hell
6
1
30
I'm quoted in this article as saying that DS bootcamps are a waste of time and the author goes on to list their favorite DS bootcamps? Strange. I'm with Rami here: learn from the community, make something, and share it with others. That's it. You don't need certifications.
This Post is from a suspended account. Learn more
1
3
27
I see we've re-entered the "your programming language has bad syntax" phase on twitter, wonderful
3
2
27
now that we've got assignment operators out of the way (a classic opener), let's knock out the rest quickly and get on with our week: - dots in variable names - for loops are slow - 0-based vs 1-based indexing - dataframes are bad actually - tidyverse = too many dependencies
2
2
7
"but can ggplot do _this_?" thankfully no, let's keep it that way
Bill the Lizard (@lizardbill) profile pictureBill the Lizard@lizardbill5 Jun 2022
Why is this even possible?
Line chart where each line is a different color gradient.

ALT Line chart where each line is a different color gradient.

8
5
76
Replying to @ogiychook
@ogiychook ggbraid v0.2.2 is now on CRAN 🎉 cran.r-project.org/web/packa…
1
1
I have, yet again, bought a domain name for a new project before making any meaningful progress on the project itself how grimly predictable
3
2
20
When I finished my grad stats program in 2017, I had no idea how to use a database. I had to teach myself, but nearly everything I found online was written for business people whose Excel spreadsheets got too big. It's been 5 years. Where do you point a data person today?
3
2
13

No more items