Every Wednesday, I publish a new issue of Bamboo Weekly, asking you questions about a real-world data set that's in the news. And every Thursday, I publish my solutions and explanations.
I've been teaching for more than 30 years, and if there's one thing I've learned it's that my solutions are often not as good as yours. Maybe your is more efficient, or elegant, or just approached the problem in a smarter way. Heck, when I'm putting together the Thursday edition, I often revise my original solution, and wonder what I was thinking.
And you know what's better than learning from me? All of us, learning from one another.
So I'm starting the Bamboo Weekly community contest, an occasional challenge to find the best solutions to that week's questions. It's your chance to gain fame and glory, while also getting better with data analysis and Pandas.
How it works
- Wednesday: I publish the questions, along with some background and the data set you need to solve.
- Thursday: I publish my solutions and explanations.
- Monday morning: Deadline for you to submit your contest entry.
- Tuesday: I publish the winners, with names and links
How to enter
Put your work in a public GitHub repository — a notebook, a script,
whatever you actually wrote — and include a link back to the Bamboo Weekly issue
it answers.
Then post the link to your repo in the comments on that week's issue. You will need a (free) Bamboo Weekly account to comment.
Post it in the comments rather than emailing me, because half the point is that
you can see each other's work. Some of the most useful things I have learned
about Pandas came from watching someone else solve a problem I thought I had
already finished. If you would rather not post publicly, email me the link
instead and I will still read it.
The three categories
Best solution to the open questions. The first two questions in every issue
are open to everyone, so this category is open to everyone.
Best solution to the members' questions. The remaining questions are for
paying subscribers, so this one is for them.
Most interesting question of your own. What would you like to learn from the week's data set? Pose a question, and then answer it. This category is open to
everyone, including free subscribers, because it needs nothing but the data. To be honest, I'm most excited by this category.
What "better" means
Any of these, and you do not need all of them:
- Faster — measurably, on the same data
- Shorter or clearer — fewer moving parts, easier to read six months later
- More idiomatic — a Pandas feature that fits the problem better than what
I reached for - More correct — an edge case I missed, or a wrong assumption in my answer.
These are especially welcome.
I will feature at least one submission in each category, and more when they
deserve it.
Explain your thinking, even briefly. A paragraph on why you did it that way is
worth more to everyone reading than another block of code, and it is usually
what decides which submissions I feature.
Who can enter
Anyone with a Bamboo Weekly account, free or paid. Free accounts can see the
first two questions and the data, which is enough for two of the three
categories.
What winning gets you
Your name, a short description of what your solution did differently, and a link
to your GitHub, LinkedIn, personal site — whatever you want people to see.
That is the whole prize for now. I may add others later.
One piece of housekeeping
By posting a submission you are agreeing that I may quote from it, describe it,
and link to your repository and your profile in Bamboo Weekly. If you would
rather I did not, say so in your comment and I will leave you out of the
write-up — you can still enter.
Questions
Reply to any Bamboo Weekly issue, or leave a comment. Both reach me.