The Price Map

One policy rate, fifty price levels. Almost every inflation chart shows a rate of change. This one shows the level: what things actually cost, and where. A national index tells you how much a basket moved; it cannot tell you that the same basket costs materially more in one metropolitan area than in another, and keeps doing so year after year.

Three sources answer that question at three different grains, and they are shown as three separate charts because they are three separate measurements — 221,489 monthly dollar prices, 3,386 annual metropolitan price comparisons, and 721,840 daily wholesale rows standing in for 1,671,751 individual quotes.

Three clocks, never merged

• Dollar prices — monthly, 1973-10 to 2025-06. This is older than the index on the prices page: it reaches this project through a mirror that was last refreshed 2025-09-24.
• The metropolitan cost map — annual, 2008 to 2023. The oldest of the three, and a structural picture rather than a nowcast.
• Wholesale quotes — daily, 2025-07-24 to 2026-07-24. The freshest, and the shortest: a rolling window of 252 trading days with no history behind it.

Nothing on this page blends the three. In particular the dollar prices are retail — what a shopper pays — and the wholesale quotes are what buyers pay at terminal markets. They are different stages of the same supply chain and must never be read on one axis.

1 · What it costs, by place

Average retail prices in dollars, not index points: 159 priced items across 73 areas, 926 series in all. Two areas can carry the same index value and completely different prices; this is the panel that says which.

2 · The metropolitan cost map

Regional price parities compare a place’s overall price level with the national average, which is set to 100. A reading of 112 means prices about 12% above the national level; 88 means about 12% below. The opening view is the most and least expensive places in the latest year, because the spread is the finding.

This is not a complete grid. Four expenditure measures are published across 158 places, but only 212 of the 632 possible place-and-measure pairs exist in the source. Most places carry just one of the four, so switching measure changes which places appear — and the count shown under the chart is always the count for the measure selected, never the total.

3 · The daily wholesale tier

Terminal-market quotes from 11 cities covering 314 commodities, daily. This is the only sub-monthly price panel on the site — and the only one that is genuinely current.

Wholesale prices are quoted per package, and the package is part of the selection rather than a footnote: a price per 40-pound carton and a price per 50-pound sack are not the same measurement, so they are never pooled into one line. 624 distinct packages are named across the source. 3,034 of the 1,671,751 quotes arrive with no package recorded at all; those are kept and labelled as unrecorded rather than dropped or assigned a likely package, they are grouped only with each other, and they are never the opening view.

What’s in here, and what isn’t

• These are three different clocks and they are never merged. The dollar prices run monthly and stop over a year before the index on the prices page; the metropolitan cost comparison is annual and stops several years before that; the wholesale quotes are daily and current but hold barely a year of history.
• The wholesale quotes are what buyers pay at terminal markets, not what shoppers pay in shops. They are a different stage of the same supply chain and must never be read on one axis with the retail prices above them.
• The metropolitan cost comparison is not a complete grid. Most metropolitan areas here carry only one of the four expenditure measures, so switching measure changes which places appear. The count of places shown is always the count for the measure selected, never the total.
• Wholesale prices are quoted per package, and packages differ by commodity and sometimes by city. Prices are only comparable within one package, which is why the package is part of the selection rather than a footnote.
• A small share of the wholesale quotes arrive with no package recorded at all. They are kept and labelled as unrecorded rather than dropped or assigned a likely package, they are grouped only with each other, and they are never the opening view — because a dollar figure whose package is unknown cannot honestly be compared with one whose package is known.
• The dollar prices reach this project through a mirror rather than as a direct collection from the Bureau, and the mirror was last refreshed well before the most recent official releases.
• Item coverage outside the national average is uneven: most places carry a small subset of the priced items, and a sparse line is a publication cadence rather than an unchanging price.

The names of the priced items and their areas are not distributed in any lookup table this project holds, so they were recovered from the source’s own series titles. Where the same area is named by two independent sources, both spellings ship as columns and the build fails rather than publish a label it cannot reconcile. Anything unrecoverable is written as the literal words NOT CAPTURED — never guessed, never filled in from a plausible neighbour. Full method: Methodology.