Business Datasets: 3 Official Sources for Public Tables

To publish a table comparing industries across counties, start with an aggregate business dataset—not a list of company names. To compare company financial statements, start with filing-level data instead.

These three official sources support different questions. All are tied to U.S. statistical programs or SEC filings; they are not worldwide business directories. The workflow below turns one source into a documented public table readers can filter and compare.

Choose the source that matches your question

1. County Business Patterns: establishment counts and payroll

Use it for: “Which counties have the most employer establishments in this industry?”

The U.S. Census Bureau’s County Business Patterns (CBP) is an annual series covering establishments with paid employees. It includes establishment counts, employment during the week of March 12, first-quarter payroll and annual payroll. Its downloadable CSV files are tabulated by geography, industry and establishment employment size. Start with the CBP dataset page; the program homepage also directs readers to record layouts and reference files.

For a first project, select one year, one geographic level and one industry level. Keep industry codes beside readable industry labels so readers can inspect both.

Interpretation limit: An annual release does not make every measure an annual average. Label CBP employment as a March snapshot, not average annual employment. The population is employer establishments, not every business, and these files provide aggregate statistics rather than a directory of named companies. See the Census program description.

2. QCEW: employment and wages by industry and area

Use it for: “How do employment and average pay compare across county industries?”

The Bureau of Labor Statistics’ Quarterly Census of Employment and Wages (QCEW) publishes employment and wage statistics covering more than 95% of U.S. jobs, with county, metropolitan-area, state and national detail by industry. See the QCEW overview.

The download page offers quarterly and annual-average CSVs by area or industry, plus single-file downloads. For spreadsheet work, start with a by-area or by-industry file: the annual versions include title columns that the single files omit, as shown in the annual CSV layout.

Interpretation limit: QCEW excludes self-employed workers and some other groups. Its wages are not simply advertised salaries: wage totals can include bonuses and other payments, and averages depend on workforce composition. These distinctions matter when labeling a pay comparison, as explained in the BLS technical notes.

3. SEC Financial Statement Data Sets: company financial disclosures

Use it for: “How do selected financial measures compare across SEC filers?”

The SEC’s Financial Statement Data Sets contain numeric information extracted from the primary financial statements in XBRL filings. XBRL is a structured reporting format that labels financial facts. Downloads are posted quarterly and preserve information “as filed.”

Each ZIP contains four tab-delimited text tables: submissions (SUB), numeric facts (NUM), tag definitions (TAG) and statement presentation (PRE). The SEC documentation explains how to join numeric facts to submissions using adsh, the filing accession number.

This is the more preparation-heavy option. Before publishing, select a measure, unit, period end, reporting duration and entity or segment scope, then document how you handle repeated periods and amendments. A download quarter identifies a filing batch—not necessarily the quarter measured by every fact. Check selected facts against the original filings: the SEC warns that the extracts may contain errors and are not substitutes for those filings.

Build a county-industry public table

A manageable starter project is a 2025 QCEW annual county-industry comparison. As of October 25, 2026, the BLS download page lists 2025 annual CSVs by area and industry.

Define the rows before cleaning

Choose counties in one state and a consistent industry level. For a three-digit industry comparison, use these filters:

• own_code = 5: private ownership, according to the ownership code table.
• agglvl_code = 75: county, three-digit NAICS industry, by ownership sector. NAICS is the North American Industry Classification System; the aggregation code table distinguishes these rows from totals and other industry levels.
• size_code = 0 or blank: all establishment sizes, according to the size code table.
• year = 2025 and qtr = A: the chosen year and annual period, as defined in the annual CSV layout.

Restrict area_fips to the intended counties. After those filters, aim for one row per county–industry–year.

Use this proposed column selection—not fabricated observations:

Columns	Purpose
area_fips, area_title	Geographic identifier and readable name
industry_code, industry_title	Industry identifier and label
own_code, agglvl_code, size_code, year, qtr	Retained scope and period fields
annual_avg_estabs_count	Annual average of quarterly establishment counts
annual_avg_emplvl	Annual average of monthly employment levels
avg_annual_pay	Average annual pay
disclosure_code	Whether the data are disclosed

These fields and definitions come from the annual CSV layout.

Validate the export

• Import geographic and industry codes as text; preserve leading zeros.
• Exclude state and national totals from a county comparison. Do not mix parent industries with their component industries in a summed total, or sum average pay across rows.
• Preserve disclosure flags. The annual layout defines N as not disclosed; do not interpret undisclosed values as observed zeros. Exclude undisclosed employment and pay from numeric comparisons, and represent them as missing in the publication copy.
• Check the chosen row key for blanks and duplicates. Keep the scope fields so apparent duplicates can be explained rather than deleted blindly. See primary keys for published CSV datasets.
• Reopen the exported file and confirm row count, codes and numeric columns.

Publish the reviewed file with context

Keep the original download separately. Accompany the public table with its source link, download date, reference year, geographic coverage, filters, units and treatment of undisclosed values. A caption such as “Private-sector employment and average annual pay by county and three-digit industry, 2025” is more useful than “Business data.”

TablePage supports CSV, TSV, XLSX and XLS uploads and generates public dataset pages with filterable tables and shareable links. Upload only the reviewed, publication-ready file—never confidential customer, employee or internal financial records.

Before sharing, inspect the resulting page as a reader: check that identifiers survived, numeric comparisons behave correctly and the source notes remain available alongside the table or on its accompanying page.