CPI-U / NATIONAL
The price level versus the inflation rate
See how a lower annual inflation rate can coexist with a higher overall price index. Pick any two published checkpoints to recompute the cumulative price-index change.
OPEN DATA WORKSHEET →THE NUMBERS. THE DENOMINATORS. THE SOURCE.
Six dated, reproducible research worksheets. Each chart reveals its raw observations, timeframe, denominator, source provenance, limits, and computation. No fictitious series, real-time data claim or politician rating.
WHAT THE MEASURE ACTUALLY SAYS
The August 2026 national CPI-U all-items annual rate was 3.4%, but the index stood at 334.980 (1982–84 = 100). A lower inflation rate is not a return to earlier price levels. Comparing cumulative change with a 12-month rate, or a metro observation with a whole state, alters the meaning. Every calculation below uses disclosed, archived input.
Open the Kennedy inflation investigation for one reporting application. The charts do not assign causal blame.
INFLATION / HEADLINE
Annual inflation rates at selected checkpoints
An annual percentage change is not a price-index level or a cumulative change since a particular presidency. Compare like-for-like 12-month observations, not an invented trend between them.
OPEN DATA WORKSHEET →INFLATION / COMPONENTS
One date, different consumer-price categories
The same economic period produced different inflation experiences for gasoline, food, energy and other goods and services.
OPEN DATA WORKSHEET →GEOGRAPHY / REGIONS
Four regions, one reporting window
Compare the Northeast, Midwest, South and West directly against each other and the national 3.4% benchmark for the SAME 12-month period.
OPEN DATA WORKSHEET →GEOGRAPHY / LOCAL
Selected metropolitan price comparisons
Contrast selected metro inflation observations while keeping place names, metro boundaries and sample-selection limits visible.
OPEN DATA WORKSHEET →CENSUS / TURNOUT
Voting totals versus turnout percentages
Election-turnout headlines change when using the voting-age population, citizens, registered voters or raw votes. Here both displayed percentages use the same CVAP denominator.
OPEN DATA WORKSHEET →HOW THIS LAB PREVENTS MISLEADING STATISTICS
Measure first
Separate totals, percentages, percentage-point differences, price indexes and model estimates. No mixing CPI and quarterly PCE.
Choose a defensible baseline
Use only the documented time points; show the exact first and last observation behind a change.
Preserve geography
National, broad regional and metropolitan figures retain distinct boundaries. No inferred state rankings.
Download and recompute
CSV, JSON, SHA-256 hashes and independent Python reproduction are included; browser controls never alter the approved inputs.
RELEVANT INVESTIGATIONS AND FIELD GUIDES
Kennedy: prices, policies and Hormuz · Before you believe a wage chart · Why statistics get revised