New: Roadmaps ordered paths through our cheat sheets and flashcards, so you always know what to study next.
Explore themSee what's new on GitHubFrom summarizing a dataset to designing an experiment that holds up to scrutiny.
A 11-step learning path. Follow it in order, or jump to what you need.
This path is for anyone who needs to read, question, and act on numbers without necessarily writing a line of code, from analysts and researchers to managers who review other people's charts. Plan on about 4 to 7 weeks at a few hours a week, moving from descriptive statistics and probability through confidence intervals, hypothesis testing, regression, and a full designed experiment. It does not cover the Python data stack or machine learning models; that applied, code-heavy depth belongs to the Data Scientist path instead. By the end you can summarize a dataset honestly, put a real confidence range around a number instead of a single misleading figure, and design a study whose conclusions would survive a skeptical review.
No prior experience needed. Start from zero.
Open this first: you can size up any dataset's center, spread, and shape today, without touching a formula you haven't seen, and every later step in this path leans on that same three-lens view.
Step 1 told you how spread out your data is; probability tells you why it landed there, the language every confidence interval and hypothesis test in this path is written in.
Go deeper on the shapes step 2 only sketched: matching a dataset to the right distribution, normal, binomial, Poisson, is what makes every interval and test downstream trustworthy instead of guesswork.
You can size up a dataset's center, spread, and shape, and put real probability language around how uncertain it is. Next up: turning that language into an actual range and a real decision.
Finish this section to unlock.
+100 XP
Turn the distributions from step 3 into an honest range instead of a single misleading number, the estimation habit that step 5's hypothesis tests build directly on.
This is where the confidence interval from step 4 turns into a yes-or-no decision: expect to come back to hypothesis tests more than once before picking the right one feels automatic, the point the field agrees is where statistical thinking actually clicks.
You can put an honest range around a number and test whether a pattern is real or just noise, and a few minutes of due flashcards keeps step 1's descriptive stats fresh while you build on it. Next up: modeling how variables relate, and handling data that isn't perfectly clean.
Finish this section to unlock.
+100 XP
Extend the hypothesis testing from step 5 to relationships between variables: fit a line, read its diagnostics, and know when a coefficient is trustworthy enough to act on.
Real datasets rarely arrive as clean as a textbook example: diagnose why values are missing before you delete or guess your way into a biased version of the regression from step 6.
You can fit and diagnose a regression model, and you know why values go missing before you delete or guess your way past them. Next up: the deeper methods, and the one full experiment that ties everything together.
Finish this section to unlock.
+100 XP
Take this if you're heading toward research-heavy or multivariable analytics work: go beyond the one-relationship-at-a-time view from step 6 to see how many variables move together at once.
Pick this up if you need to argue that one thing actually caused another, not just moved with it: these frameworks justify a causal claim from data you can't randomize, the exact problem step 11's designed experiment sidesteps by randomizing in the first place.
This one's for you if a little Python doesn't scare you: get a second lens on step 5's hypothesis tests, where instead of a p-value you get a full posterior distribution over what's actually likely true.
Put the randomization and hypothesis testing from step 5 to work in one designed study: plan it, run it, and analyze it well enough that the conclusion would survive a skeptical review.
You can describe a dataset honestly, put a real range and a real test behind a claim, model how variables relate, and design an experiment whose conclusion would survive a skeptical review. That's the statistical literacy the Statistics for Data Work badge certifies.
Finish this section to unlock.
+100 XP
Finish every required step, at least 70% of them genuinely done (not skipped), to earn this badge and 500 XP.