BLS Consumer Price Data Tools (Nonalcoholic Beverages CPI)

BLS Consumer Price Data Tools (Nonalcoholic Beverages CPI) is the Soft Drinks & Non-Alcoholic Beverages catalog's record over the consumer price line: the item categories of the US Consumer Price Index that cover what households pay for drinks without alcohol.

The structure is a family tree, not a flat list. At the top sits SAF114 - nonalcoholic beverages and beverage materials - the aggregate most dashboards quote. Underneath, SEFN covers juices and nonalcoholic drinks and splits into SEFN01 carbonated drinks, SEFN02 frozen noncarbonated juices and drinks and SEFN03 nonfrozen noncarbonated juices and drinks, while SEFP covers beverage materials including coffee and tea. Each code combines with an area code and an adjustment prefix to form a full series identifier - CUUR0000SAF114 reads as not-seasonally-adjusted, US city average, nonalcoholic beverages - and the same category exists again under the CUSR prefix, seasonally adjusted. Observations are index values on the 1982-84=100 basis; the aggregate runs from January 1947 through July 2026, and each year closes with an annual-average row coded M13.

For a beverage business this is the shelf-price environment in official numbers: the series your revenue deflates against, your pass-through story tests against and your escalation clauses cite. Get a sample of this dataset cut to the categories, years and geographies your models touch.

A monthly observation and the series header it belongs to, exactly as they arrive:

series_id : CUUR0000SAF114
year      : 1947
period    : M01
value     : 11.4

series_id   : CUUR0000SEFN01
title       : Carbonated drinks in U.S. city average, all urban consumers, not seasonally adjusted
base_period : 1982-84=100

The first record is the oldest observation in the aggregate: in January 1947, nonalcoholic beverage prices stood at 11.4 percent of their eventual 1982-84 level - a number that makes later decades of beverage inflation legible in one glance. The second is a series header, and its presence is the point: rows arrive self-describing, with the plain-language title and the index basis carried beside the identifier, so nobody has to keep a private lookup table to remember what CUUR0000SEFN01 means. Identifiers encode the adjustment status and the geography, titles state the category in English, and the value column stays a pure number on a stated base.

Five fields define every observation, each definition checked against the record during research rather than inferred from prose. The identifier does the heavy lifting: adjustment prefix, area code and item code stack into one string, which makes series_id x year x period a natural key that joins cleanly against your own price and revenue tables without a crosswalk.

Four more columns appear on parts of the corpus and therefore fold under additional fields on request: title carries the plain-language series name on headers, base_period states the index basis, area_name decodes the geography behind the area-code segment, and item_code isolates the beverage category behind the trailing segment. Say which ones your pipeline expects when you request a sample.

Geography - the US city average leads every series, and regional and area indexes hang off the same identifier scheme, so a national view and a regional cut share one schema instead of forcing a second ingestion path.

Temporal - the nonalcoholic beverages aggregate starts in January 1947 and runs through July 2026 as of extraction, roughly 950 monthly observations on a single series. One honesty mark matters before modeling: October through December 2025 carry an X footnote marking them unavailable after the 2025 lapse in appropriations, so the recent window has a visible seam rather than a silently interpolated one.

Granularity - one index observation per item category per month, with annual averages arriving as M13 rows and half-year averages available alongside. Nothing is resampled upward to pretend a monthly panel is daily; you choose the cadence that matches the decision.

API, files, or your warehouse. Daily, weekly, or hourly.

Your cadence is your call regardless of how the underlying series behaves - a monthly index still lands on your schedule, whether that is a nightly warehouse load that keeps pricing dashboards fed or an on-demand pull for quarterly board material. Nobody babysits the pipeline.

Every delivery ships the complete field dictionary above unchanged, the sample rows for validation and a coverage profile mapped to the categories, geographies and years you named, so the extract arrives pre-cut rather than as a pile for your team to sort.

• Revenue deflation and real growth - converting nominal beverage sales into volume terms needs a price deflator matched to the category mix, which is exactly what the carbonated-drinks and juices indexes provide.
• Pass-through measurement - when input costs move, the consumer indexes show how much reached the shelf and how fast, turning a pricing-decision debate into a measurable lag.
• Contract escalation and indexation - CPI-linked clauses in vending, foodservice and retail leases require a clean monthly series with published history and a stated basis.
• Relative-price and substitution studies - the carbonated-versus-juice trajectories quantify the relative-price drift the reformulation wave rode in on.
• Staples-inflation forecasting - a nine-decade monthly panel is training and benchmark material for consumer-staples inflation nowcasts.
• Margin narrative for strategy and IR - official index numbers hold up in board decks where vendor anecdotes do not.

Market researchers and consultants (relevance 3/3) get an official shelf-price backdrop for every category conversation, going back to 1947; see market researchers use cases. Data scientists and ML engineers (3/3) get a flat five-column schema with a natural key that drops straight into feature pipelines; see data scientists use cases. Developers building data products (3/3) get one identifier scheme across every category and geography; see developers builders use cases. Competitive intelligence and product teams get the consumer-price context behind category moves; see competitive intel product teams use cases. Investors and quant researchers get a long monthly staples-inflation factor, and journalists, academics and students get citation-grade provenance on every number.

Within soft drinks and non-alcoholic beverages data, this record owns the consumer price line: what households pay for drinks, monthly, since 1947, in dollars-of-the-day deflated to a stated basis. The neighbors own different jobs. UNESDA - Soft Drinks Europe Market Data owns European volumes and per-capita consumption - industry counts, yearly, across 29 markets, not US prices, monthly. USDA ERS Sugar and Sweeteners Yearbook Tables owns the input side: what producers pay for sweetener and how much moved, upstream of the register. CDC NHANES - Dietary Intake & Beverage Consumption Data captures what people report drinking, survey-style. If the question is "what happened to drink prices at the shelf, and when", this is the set that answers it. The head-to-head argument continues in vs WHO Global Health Observatory Indicators.

Four things worth knowing upfront.

First, these are indexes, not dollar prices. Values are expressed against the 1982-84=100 basis, so the analytic quantity is the percent change - month over month, or year over year - not the level. A reading of 11.4 in January 1947 means prices sat at 11.4 percent of their 1982-84 average; budget models that treat the number as a price per litre will produce confident nonsense.

Second, the 2025 seam is flagged, not hidden. October through December 2025 carry an X footnote marking them unavailable after the lapse in appropriations. Treat flagged-missing as missing; zero-filling turns a funding gap into a fake price collapse.

Third, every category comes as a pair. Seasonally adjusted and unadjusted versions coexist, distinguished in the identifier prefix. Year-over-year charts tolerate the unadjusted twin; month-over-month models usually want the adjusted one - pick per use case and stay consistent.

Fourth, start dates vary by category. The aggregate reaches January 1947; where a narrower cut such as frozen noncarbonated juices begins is a per-series question. Name the categories you need and the sample states each series' actual start date, with the field dictionary and coverage profile attached.