Some Platypus Economics videos come with a companion worksheet, built in partnership with Stata. Watch the video for the intuition, then open Stata and work through the data yourself — each worksheet pulls real numbers, runs a handful of commands, and lets the pattern emerge in your own hands. Free to use in your classroom or on your own.
New to this? You'll need Stata and a free FRED API key — grab one at fredaccount.stlouisfed.org/apikeys. Each worksheet's first step shows you how to set it up (importing FRED data needs Stata 15.1 or later).
Does higher inflation mean more volatile inflation?
Pull decades of inflation data from FRED and measure it two ways — the average rate and how much it bounces around. See for yourself whether the two move together, and what that means for planning a life around prices you can't fully predict.
The building blocks of descriptive statistics — the mean as a measure of the average and the standard deviation as a measure of dispersion — plus computing growth rates as percentage changes and comparing the spread of a variable across groups.
Paired with the video “You Can't Stop Inflation, But Here's How to Minimize the Pain”
Does gas rise like a rocket and fall like a feather?
Line up weekly gas and oil prices from FRED, then run one regression to test a famous asymmetry: the price at the pump shoots up like a rocket when oil rises, but drifts down like a feather when oil falls. Measure the pass-through on impact and eight weeks out.
Regression as the workhorse of econometrics — here a distributed-lag model with several lags of one regressor, dummy and interaction terms that let an effect differ by direction, log changes as approximate percentage changes, and summing coefficients into a cumulative effect.
Paired with the video “You're Not Imagining It: Gas Prices Really Do Fall Slower Than They Rise”
Did tariffs on China reduce imports, or change their source?
Pull monthly import data from FRED for China and all eleven ASEAN members, then smooth it into rolling 12-month totals and watch a 25-year relationship flip sign. Before the 2025 tariffs, America bought more from ASEAN whenever it bought more from China; since then, every dollar lost from China shows up somewhere in Southeast Asia. Run one regression per era and measure the flip yourself.
Working with time series — building a monthly date, computing rolling 12-month totals that strip out seasonality, and aggregating series across countries — plus splitting a sample at a structural break, reading scatterplots with fitted lines, and interpreting a regression slope that flips from complements to substitutes.
Paired with the video “The Muddled Thinking Behind Trump's Tariffs”
Pull quarterly data from FRED on what corporations pay their workers and what they earn, then compute labor's share of income two ways — out of every dollar earned, and out of every dollar left after depreciation. Both measures have sunk to all-time lows. Run the numbers yourself and decide: depreciation changes the number, but does it change the story?
National income accounting made concrete — gross versus net value added, and why depreciation isn't income to anyone — plus building shares from raw dollar series, checking how a conclusion holds up when you change the denominator, and using summary statistics to judge whether today's value is an all-time low.
Paired with the video “The Great Shift from Workers to Owners”
What a $20 Burrito Can Teach Us About Affordability
Pull sixty years of wages and consumer prices for twelve major economies straight from the OECD, then line them up country by country. From the 1970s inflation spike to the 2020s, wage growth and inflation rise and fall together almost everywhere. But "together" isn't the same as "one-for-one" — run the regression yourself and test whether a percentage point of inflation buys workers a full percentage point of wage growth.
Working with panel data — importing from a web API, merging two sources on country and year, and using tsset so lags respect country boundaries — plus building growth rates from index levels, reading small-multiples charts to spot a pattern across countries, and testing a specific hypothesis about a regression coefficient (b = 1) rather than just whether it differs from zero.
Paired with the video “Every Ingredient in This Burrito Is a Policy Failure”
Pull the BEA's regional price parities straight from bea.gov, then paint all fifty states — plus one gate-crasher — in five shades from bargain to very pricey. Prices run about 27% higher in California than in Arkansas, which is why a $75,000 offer in Michigan can quietly beat an $85,000 offer in California. Draw the map yourself and find out which shade you're living in.
Making price indexes concrete — what "US = 100" means and how to turn a nominal salary into real purchasing power — plus downloading and unzipping government data without leaving Stata, taming a messy CSV whose year columns arrive as v-names, installing community add-ons from SSC, and drawing a choropleth map with hand-picked bins and a legend written in plain English.
Paired with the video “What Would It Take To Eliminate Poverty”
Pull the jobs report straight from the Bureau of Labor Statistics the morning it drops, with no waiting for the news to tell you what it means. You'll build your own summary table, separate this month's noise from the three-month trend, and check whether unemployment moved for good reasons or bad. Then you'll find the surprise. Since January 2025, health care and social assistance has added nearly a million jobs while the rest of the economy has gone backwards, and more than nine in ten net new jobs have gone to women. Run it yourself and write the headline before the pundits do.
Reading the two surveys behind every jobs report: the payroll survey that counts jobs and the household survey that measures unemployment and participation. You'll also learn why monthly changes are noisy and a three-month average helps, and how a falling unemployment rate can be bad news. On the Stata side: registering for an API key and pulling data straight from a government API, installing a community command from GitHub with net install, and taming messy downloads where missing months arrive as "-" (destring, force). You'll also stack and reshape data from long to wide, compute changes with lagged values, and pick out specific rows with if and _N.
Paired with the video “What September's Employment Numbers Really Say”
Health care and social assistance has quietly become the engine of the American job market. Pull the payroll data from FRED and you'll see that its share of all jobs rose from about 8% in 1990 to around 15% today, and that it accounts for nearly half of all jobs added since COVID. Then test the stories everyone tells about it. Did the pandemic leave a hole that took years to fill? Was the 2022–24 hiring surge really people catching up on missed checkups, or something else? And is that post-COVID echo finally over? Run the numbers yourself and see what the data says.
Students track the long-run structural shift toward care work and think about why richer, older societies spend more of their income on health, from income elasticity to Baumol's cost disease. They pull several BLS payroll series at once with import fred, give them readable names with rename, and set up monthly time-series data with mofd() and tsset. They compute a sector's share of total employment and use D. to turn job levels into monthly gains. With regress and predict, they fit a pre-COVID trend and project it forward to measure the pandemic hole and the catch-up, then compare average monthly gains across periods with summarize and if conditions to judge whether hiring has returned to normal. Finally, they use list to compare subsectors and test a popular explanation against the evidence.
Paired with the video “Why America Spends So Much On Health Care”
Pull the latest inflation and wage numbers straight from the Bureau of Labor Statistics, on the day they’re released. You’ll chart headline and core inflation over the past decade, from the 2022 peak to today. Then you’ll put wages and prices on the same footing to see whether an hour’s work buys more or less than it did a year ago. The surprise is that over the past year prices rose 3.7% while wages rose only 3.0%. The typical worker took a real pay cut, and you’ll measure it yourself.
You’ll learn the difference between headline and core inflation and why economists strip out food and energy, how to turn a price index into a year-ended inflation rate, and how comparing cumulative wage growth with cumulative price growth reveals what’s happening to real wages. Along the way you’ll pull data directly from the BLS API with getbls, clean up a month the government never published using destring, force, stack and reshape series into a tidy monthly time series, use the lag operator L12. to compute 12-month changes, index series to a base month using summarize and r(mean), and plot and export graphs with twoway line.
Paired with the video “The American Economy Just Got More Bad News”
Watch 
