pandas
Grouping, joining, reshaping — learned on your data, in your hands.
// how_a_data_science_session_works
Anatomy of a data-science session.
Bring a CSV you actually care about — from your job, your thesis, or a field you love — and we run the whole pipeline on it live: clean, explore, chart, conclude. You leave with a finding you can show, not an exercise you’ll forget.
- subject: data_science
- format: end_to_end
- data: bring_your_own
- output: portfolio_ready
Your first session is free. No card.
// your_data_not_toy_data
The dataset is yours. So is the finding.
Tutorial datasets teach you tutorials. Your own data — with its missing values, weird columns and real stakes — teaches you data science. Every session runs the honest pipeline: from messy file to a one-line claim a chart can defend.
- We build on your data. Bring a CSV from your job, thesis, or a dataset you care about and we run the whole pipeline on it — clean, explore, visualise, conclude — so the work is portfolio-ready, not a toy.
- pandas, in your hands. You learn the grouping, joining and reshaping moves on the live data in front of you, and we save the notebook so the analysis is reproducible later.
- A result you can show. Each project ends with a chart or a short written finding you can drop into a report or a portfolio — work you built yourself, which is exactly why it sticks.
analysis.ipynb
// the_60_minute_hour
What a real data science hour looks like.
The same deliberate rhythm as every session — tuned to how data science is actually learned.
0:00
Pick the dataset & the question
Yours if you have one; a real public dataset matched to your field if you don’t. We phrase the question as a claim the data could prove wrong — that’s what separates analysis from chart-making.
a falsifiable question0:05
First contact with the data
df.head(), df.info(), the missing values, the column that’s secretly a string. Before any analysis we establish what’s actually in the file — most wrong conclusions are born in this skipped step.
know the file first0:15
Clean, reshape, explore
The core of the hour, with you typing: the pandas moves — filtering, grouping, joining, reshaping — applied to live data whose quirks you have to handle for real, not a sanitised exercise.
real moves, real mess0:40
Chart & claim
You produce the figure yourself and write the one-line finding it supports — then we stress-test the claim together. Could the pattern be seasonal? An artifact of missing data? Now you check.
a claim the chart defends0:52
Reproducible hand-off
The notebook is saved runnable top-to-bottom, the chart and finding are yours for a report or portfolio, and the written summary lands by email with what to practise next.
portfolio-ready output
// the_toolkit
Four tools. Nothing to install.
Everything runs in the browser, and everything we make stays yours afterwards.
matplotlib & friends
Charts built to defend a claim, not decorate a slide.
Your own CSVs
Job, thesis, or a field you love — real stakes make it stick.
Reproducible notebooks
Saved runnable top-to-bottom, so the analysis survives the session.
// honest_answers
Data Science questions, answered straight.
The things people actually ask before their first data science session.
Can I bring data from my job?
Yes, and it’s often the best material — you know its context and you care about the answer. I treat anything you share as confidential. If your employer prefers nothing leaves the building, we mirror your problem on a public dataset with the same shape, and you re-apply the notebook at work.
I only know Excel — is it too early for this?
Not at all; Excel-to-pandas is one of the most common journeys I teach. Every move you already make — filters, pivot tables, VLOOKUP — has a direct pandas equivalent, and we build the bridge from the tools you know rather than starting from zero.
Do sessions cover statistics?
Yes, exactly when the analysis needs it. The moment a finding depends on a statistical idea — significance, correlation versus causation, distributions — we step to the whiteboard, work it through, and return to the code. Statistics taught next to your own data is statistics you remember.
What if I don’t have a dataset?
Then we pick a real public one matched to your field or interests — sports, climate, finance, health, whatever you’ll genuinely be curious about. The only rule is no toy datasets: the mess is where the learning is.
What do I actually walk away with?
A notebook that runs top-to-bottom, a chart with a finding you can put in a report or portfolio, the written session summary, and — over several sessions — a body of analysis work that’s genuinely yours to show.
// start_here
Bring a real data science problem.
The fastest way to understand a session is to have one. Pick the thing you’re actually stuck on and we’ll work it together — no slides, no script.
Your first session is free. No card. Cancel any time.