Find SIC 2007 codes for an industry. Everyday words (AI, coffee shop, barber, e-commerce) first return ready-made bundles: approximate code sets, with what they include and miss. Pass a bundle's id in sic_codes and say it is approximate. Then codes for that activity, then codes whose description contains the words (whole words, up to 15, sorted by code, not relevance: keep only the ones for the activity asked about), then their 3-digit groups. For a broad industry ("cleaning"), pass the group code (e.g. "812") so every code in it is counted. Name the codes you used in your answer.

Find areas by name, matching from the start of any word: regions and nations first (London means all 33 London council areas; the City of London is one small council area), then local authority districts (England, Wales, Scotland). A county isn't an area here, so a county's name may find one district in it: check the name. Northern Ireland has no area breakdown.

Count companies (private limited, PLC and LLP) registered in the period with these primary SIC codes that are still live in the snapshot, or with include_closed the companies in this data registered, including those dissolved by the snapshot date (use this to compare periods fairly). Area answers leave out formation-agent postcodes.

One-call brief: is this industry growing faster or slower in this area than across Great Britain? Compares a period (default the latest four quarters) with the same quarters a year before, or the period just before: the area's count, Great Britain's, the area's share of GB with a significance test, how concentrated the industry is here (location quotient), and the quarterly trend. Formation-agent addresses are left out on both sides. If you picked this area from rank_areas_against_great_britain, quote the ranking's verdict instead: this tool tests one area on its own. To say which of two areas grew faster, use compare_two_areas.

Head to head: which of two places has more new companies in this industry, and which grew faster? Gives each area's count for a period (default the latest four quarters) and the earlier one, a test of whether one has more, and a test of whether one area's share of the two together changed (that is, one grew faster, or fell less if it fell). Both tests allow for rounding. Formation-agent addresses are left out.

Hotspots: in which council areas did this industry's share of Great Britain's new companies rise (or fall) most, in a period (default the latest four quarters) against an earlier one (see against)? Every area is tested, so the bar for "stands out" is raised to allow for testing hundreds of areas at once. Only areas that stand out are listed, strongest evidence first. Areas without at least 10 companies for certain in both periods aren't tested. Counts include companies dissolved since, so an area isn't made to look like a hotspot by closures in the older period. This doesn't say which area grew fastest: an area missing from the rows may have grown as fast on less evidence. To check one named area, use compare_area_with_great_britain.

What's growing here? New companies in every 2-digit SIC industry division (plus no activity stated) for a period (default the latest four quarters) against an earlier one, with a test of each change and, for an area, the same test across Great Britain. Every division is tested, so the bar for a clear change is raised for testing them all at once. Rows are in SIC order, not by size. Counts are shown only for divisions with a clear change, and even those rounded counts don't reliably show which division is bigger or moved more: don't rank divisions. Formation-agent addresses are left out of area answers.

How crowded is it? Count companies (private limited, PLC and LLP) with these primary SIC codes that are live (not dissolved) in the snapshot, whatever year they were registered. Live includes companies that don't trade. Rows split it into those registered before 2021 and since. Area answers leave out formation-agent postcodes.

Were more companies registered or dissolved? In one place (default the UK) and period (default the latest four quarters), compares companies registered with companies dissolved, whatever year those were registered, and tests which was higher, allowing for rounding. A period under a year gets the counts but no verdict, because closures come in weekly batches. It can't show whether closures are speeding up. Area answers leave out formation-agent addresses.

Where are most of these companies? The council areas (or regions and nations) of Great Britain with the most companies registered with these codes in a period (default the latest four quarters), largest first, each with a likely range and a likely rank. Size, not growth: for where an industry is growing, use rank_areas_against_great_britain. Formation-agent addresses are left out.

Which industries are biggest here? The 2-digit SIC industry divisions (plus no activity stated) with the most companies registered in a place in a period (default the latest four quarters), largest first, each with a likely range and a likely rank. Size, not growth: for what is growing, use breakdown_by_industry. Area answers leave out formation-agent addresses.

Of the companies registered in the period with these primary SIC codes, what share were not dissolved within a year? For one area, compares it with the rest of Great Britain with a significance test; for 'UK', gives the national rate and lists the council areas whose rate passes a test for a difference (the bar is raised for testing hundreds of areas). Only quarters whose companies have all had a full year are allowed.