Thoughts on data science, software engineering, and technology
As I started to think about my meme-generating agent, I figured that I might want to allow it to ask clarifying questions. That makes it not just a multi-turn conversation (which just requires maintaining a history) but more like a ReAct agent with a "human in the loop" tool for …
As my next project, I thought it would be fun to use AI to generate memes. For many years I have used imgflip to generate memes. And they have an API! Unfortunately, they don't have an official Python library (that I could find, I found a bunch of one-GitHub-star unofficial …
"How do you know if someone does Crossfit?"
"They build a RAG system about Crossfit!"
As my next project, I wanted to build a non-trivial RAG system. To do that, I needed a non-trivial dataset.
Every day for 20+ years, Crossfit.com has published a Workout of the Day ("WoD …
A few years ago we went on an Alaska cruise, and we sprung for a cabin in the private/exclusive part of the ship (which is the way to go!) which had its own restaurant, its own lounge, and so on. The lounge had board games, one of which was …
A long long time ago I took a graduate course in Econometrics. We spent the first month or two deriving linear algebra equations and proving facts about Best Linear Unbiased Estimators and Consistency and Heteroscedasticity and so on. We were locked in. Then we had our first exam and one …
I have started several novels, although I have never finished one. Can AI maybe help with that?
One way of writing a novel I have tried is the Snowflake method, so named because of its fractal nature: start with a one sentence summary, expand it to a paragraph, expand that …
The world is getting into Vibe Coding, so I will too. As a first attempt, I asked Claude Code to redesign my website, which had grown a bit stale. (OK, a lot stale.)
I think it's an improvement.
Next time: some actual code.
A while back I had the idea that 9yo and I would program a bunch of simple games together and she'd get interested in coding. Well, we did program a bunch of simple games together:
Next @joelgrus’ book: “100 ways of writing fizz buzz in Python”
— Rocco Meli ⌬ (@RoccoMeli) June 13, 2019
I've been quarantined for the last several months for reasons that are too boring to go into, which somehow gave me a lot more time to write. And so I am delighted to …
I am thrilled to announce that the second edition of Data Science from Scratch is now available! (buy from Amazon or your other favorite bookstore, or read on Safari, or get a PDF from ebooks.com it looks like.)
It's been almost exactly four years since the first edition came …
I give talks, and this fall my talk has been to live code a deep learning library (and use the resulting library to solve Fizz Buzz).
You probably didn't get to see my talk, which is why I made a video of it:
Please enjoy!
interviewer: Welcome, can I get you coffee or anything? Do you need a break?
me: No, I've probably had too much coffee already!
interviewer: Great, great. And are you OK with writing code on the whiteboard?
me: It's the only way I code!
interviewer: ...
me: That was a joke.
interviewer …
Long-time readers may recall that last year I wrote a blog post about the mathematics of Spot It. (For those who don't recall, Spot It is a game where you have a deck of cards, each of which has 8 pictures on it, where through the magical mathematics of finite …
When I worked at Farecast we had a giant TV in the lobby, and up on that TV was an image of a globe, and on that globe were animated paths that (as far as you know) corresponded to flight searches that people were doing on the site. It was …
OK, so last time we built a web service that responds to GET requests with random (bogus) elementary school science questions. In this third (and last) installment, we'll create a single-page quiz app that consumes the service.
Here's a rough stab at a wireframe:
On the …
Last time we collected and processed the data for generating stupid fake elementary school science questions and answers. The important parts to remember are
questions.json and answers.json
containing transition dictionaries mapping each word to an array / list of
possible following …As I'm not working right now* , I have a surfeit of time to hack on stupid things. In particular, it seemed like a good idea to hack together a stupid data product.
Inspiration (if you can call it such) struck when the Allen Institute released a …
The readers of my book have been clamoring for an index of functions,
so that -- for example -- when someone sees me use vector_mean on page 200 they can
easily figure out where to find its definition.
It was easy enough (if tedious) to go through the book and create a …
[The fourth in an (at least) 6-part series, all code on GitHub as always.]
I know, you're thinking, "I've already read three parts of this series, and I haven't heard one …
[The third in an (at least) 6-part series, all code on GitHub as always.]
Lambda also allows Python functions, although only Python 2.7. In practice, this won't affect us much …
[The second in an (at least) 6-part series, all code on GitHub as always.]
AWS has a recent-ish Lambda product, which lets you upload functions and then have them run on …
[The first in an (at least) 6-part series, all code on GitHub as always.]
Like most of you, I've long dreamed of making a Twitter bot. And also like most of …
(This is a blog post version of my PyData Seattle talk, slides and code at the link.)
In my last post
If you've ever done a tech interview, you're probably familiar with the Fibonacci sequence:
1, 1, 2, 3, 5, 8, 13, ....
where each number is the sum of the previous two. A relatively simple (and relatively overused) interview problem is to write a function that returns the n-th Fibonacci number …
Last weekend we went to a party where one of the other attendees brought Spot It! Frozen for her kids. It's a simple game with circular cards, each of which has 8 pictures in it, most of them Frozen-themed.
The setup is that any two cards in the deck …
I am super-excited to announce that the book I've been working on for more than the last year, Data Science from Scratch: First Principles with Python is finally available! (buy from O'Reilly, use discount code AUTHD to save some money) (buy from Amazon).
My experience learning and teaching data science …
2014 was a crazy year, mostly because I took on two very large projects either of which would have made for a pretty crazy year.
I am writing a book. (Yes, another one.) Actually, the first draft is done, and I am right now revising it based …
(You might want to read Part 1 first.)
When last we left off, we'd built a model using shirt colors to predict boy-ness / girl-ness.
Our second attempt will involve the shirt images themselves (sort of). For our purposes, computer images are made up of pixels, each of whose color is …
Before I was a parent I never gave much thought to children's clothing, other than to covet a few of the baby shirts at T-Shirt Hell. Now that I have a two-year-old daughter, I have trouble thinking of anything but children's clothing. (Don't tell my boss!)
What I have discovered …