Data scientist at Klick (@klickhealth); previously, medical biophysics at the University of Toronto (@MBPatUofT), and some light astrophysics before that
Joined January 2014
- Tweets 1,131
- Following 676
- Followers 219
- Likes 3,913
88 Photos and videos
3 Dec 2022
Yes, and generating recipe sagas just got much easier and more fun.
3 Dec 2022
The most underrated value add of ChatGPT:
The elimination of recipe sagas
11 Oct 2022
Comparing datasets by plotting them in a Jupyter notebook, opening each plot image in its own browser tab, and quickly flipping between them to pick out differences using the gestalt principles of visual perception.
11 Oct 2022
📢 We're doing our first NormConf sticker giveaway (sponsored by @stickerspace_) 🎉🎉🎉
QUOTE TWEET this with your answer to "what's the most manual data task that you won't or don't automate and why?"
Winners will be selected on 14 Oct 2022 by a script or something🤷
4
4 Oct 2021
Around $3M and counting wolframalpha.com/input/?i=fa…
ALT A screenshot of a Wolfram Alpha result showing that Facebook's net income is about $1235 US dollars per second.
We’re aware that some people are having trouble accessing our apps and products. We’re working to get things back to normal as quickly as possible, and we apologize for any inconvenience.
4 May 2021
What business problem are we really trying to solve here?
Peter Leimbigler retweeted
2 May 2021
a large image of hundreds of flocking birds contains enough information to create an animation of a single bird flying
21
372
4,236
23 Apr 2021
Great thread which reminds me of @vboykis' work to make implicit knowledge explicit, required reading for all data roles IMO: veekaybee.github.io/2021/03/…
My answer: before starting analysis, ask yourself and your client "why are we doing this?" and agree on your business questions.
1
1
23 Apr 2021
After you have determined what you're trying to answer with data, sketch (as in freehand draw) your idea of analytic results that would really impress your client, then show them and refine as needed. These sketches are your expected output to work backwards from.
1
1
23 Apr 2021
Based on questions and expected output, list all data you think you'll need: short description, rough size, access cost, aggregation level (e.g., total regional sales vs. individual transactions), date range & frequency, update frequency, comments, ...
20 Apr 2021
20 Apr 2021
#USSTRATCOM Posture Statement Preview: The spectrum of conflict today is neither linear nor predictable. We must account for the possibility of conflict leading to conditions which could very rapidly drive an adversary to consider nuclear use as their least bad option.
Peter Leimbigler retweeted
19 Apr 2021
How it started: How it's going:
11
569
3,506
16 Apr 2021
This will help about as much as tossing an ice cube into an active volcano.
15 Apr 2021
As of April 16, 12:01 a.m., recreational camping on Crown land will be prohibited to help stop the spread of #COVID19.
Remember to follow the advice of public health officials and stay close to home.
ontario.ca/page/recreational…
3
16 Apr 2021
Ontario's misguided COVID restrictions aside, it's infuriating to watch Ford and his entourage performatively don a mask to listen to questions in silence, only to remove that mask to speak a long-winded reply. This is the exact opposite of effective masking.
1
1
8 Apr 2021
Tech companies: we maximize customer acquisition and retention by using advanced analytics to deliver tailored, data-informed messages to the right user through the right channel at the right time and cadence.
LinkedIn: the emails will continue until signups improve
ALT A screenshot of my gmail inbox shows a series of repeated promotional emails for LinkedIn Premium which all start with the phrase "Peter, thanks for being a valued member" received over the past five months. Some but not all of the emails are unread.
1 Apr 2021
For April Fool's, recipe websites should move the usual SEO-gaming text below the recipe itself. Maybe even prank us with lean, accessible web design that gives more space to the recipe than to ads and doesn't turn my CPU into a fast-boil range element. That'd really get me good!
1
30 Mar 2021
A compelling example of the underrated power of AI/ML: Add Interns/Manual Labour.
30 Mar 2021
After 17 years, we finally “cracked” a $100M churn problem at PayPal. Zero fancy tech. Just a spreadsheet, some simple SQL, and a physicist named Ben. 👇🏼
24 Mar 2021
We should all use this new feature to DM Slack employees to ask them to remove said feature.
2
19 Feb 2021
The fact that $202,102.16 looks like it came from 2021-02-16 makes this better, yet also horrifically worse.
18 Feb 2021
Apparently I paid the gas bill for most of Houston this month
Peter Leimbigler retweeted
27 Jan 2021
Pandemic Day 25: I made bread :)
Day 95: I sure do miss my friends
....
Day 310: The White House appears to be under the control of a shirtless man in a Viking helmet
Day 330: Reddit’s coordinated attack on Wall Street is going as planned
492
87,824
452,625
6 Jan 2021
Democracy system status page: "something's not quite right"
