Paris, France

What the business base is made of, what it is specialised in against the national pattern, and what that mix implies. This is the part a development office would read first.

Composition, not health. geonet records what trades exist and where; it does not record formation, closure, employment or revenue, so none of those appear here. What makes this economics rather than a directory is the NAICS coding: it lets the local mix be read against the BLS survival curves and against the 323 industry groups this research has screened.

What this city is specialised in

Location quotient is this city's share of a trade divided by the same share across 1 FR metro tables. Above 1.25 is the conventional threshold for a local specialisation.

The census: the only figures on this page that count people rather than businesses, and the three cohort readings they make possible.

Who lives here and the languages business runs in: population and age from the statistics offices, and the share who can hold a conversation in English and other languages from the EU’s 2024 language survey. Each figure says whether it covers the city, the region or the whole country.

Age: under 15 11.7% · 15–64 69.5% · 65 and over 18.8% (city (NUTS 3 FR101), 2025) Eurostat

Not established: Paris household count: the Eurostat census figure (1,392,973) looks too high for the population and has not been checked against INSEE.; France's census has no language question; no city language shares found.

A community appears in the business register before it appears in prose — a national church, a society, a credit union, a newspaper. Those are counted as institutions, and they are what suggests residents. A restaurant serving a cuisine is counted separately as trade, because every city has those and they say nothing about who lives nearby. The ranking is by institutions.

Streets, not neighbourhood names. The data knows that Greek institutions sit on one street and Greek restaurants on another; it does not know the word "Greektown", so it does not use it.

Every business on file, grouped by the sector it trades in. This is the economic-development view: where the 11,041 businesses actually are, and the specific trades inside each sector.

Brands from a register of 2,024 franchise systems, matched into this city by name. A high count is a signal in both directions: the territory works, and it may be taken.

30 of these brands are a single common word — Metro, Avenue, Budget — so their counts include businesses that merely share the word. They are marked rather than dropped.

Four unrelated jobs share the name: selling existing homes, owning and letting them, managing them for someone else, and building new ones. For anyone weighing a move or a licence these are the counts — and the ownership question goes to /property, which answers it building by building.

Real estate is a local job market in its own right, and four unrelated trades get the same name. Counts are businesses filed under each trade in the register, which lists a brokerage once however many agents work there — so this measures firms, not licensees.

The Ultimate 411 half of local: everything where the customer does not have to come to you. Each service opens a page on Ultimate 411 to request a quote or book a time.

A shop needs the customer to walk in. These 186 services do the opposite — the mobile mechanic, the mobile groomer, the bookkeeper, the junk haul — and they are the ones a person can start without a lease. 0 of them already match a trait filed in this city's register.

The 186 Service=1 rows in search.local: work on the go, mobile services, where the provider travels to the customer. "present" counts businesses already filed under that exact trait name; a blank is a coverage gap, not proof of absence, because the two vocabularies were built separately.

wander.schools is a post-secondary table — universities and colleges. K-12 public and private schools are not in it, and only 7 of its 1,686 rows carry coordinates, so these are listed and not mapped.

What a person moving a parent here would ask first. The register files clinics and pharmacies; it has no category for a hospital at all, so hospitals are matched by name and the limit is stated rather than hidden.

geonet carries no 'hospital' trait, so these are matched on the business name. A distance-to-hospital figure needs a source that states hospital locations.

Companies in this city that have raised money, and the careers pages of employers here, from the /employers research and the company database. For someone looking for work here, this is the shortlist.

Funded here — 6 companies have raised money

A fresh round usually means hiring. Largest raise first; the amount links to its source where one is on file, and the currency is as the source wrote it.

Careers pages — 52 more employers here

Each link goes to the company's own careers page or official job board, not a third-party listing site. Largest local staff first.

Also based here — 3 researched companies

Companies and software vendors from the market research whose home in the company database is this city, with the market they were researched in. No careers page is on file for these yet.

Named in full, because the gaps are the part most likely to be misread as an answer.

• No census figures on this page: this region is not mapped to a census subdivision in local/census-map.json, so the community counts here are business-register institutions, which is a different measurement and cannot substitute for a population.
• wander.schools is now readable, but it holds universities and colleges only and 1,679 of its 1,686 rows have no coordinates. K-12 public and private schools need a separate source — in Ontario the Ministry of Education school finder, in British Columbia the provincial school list. wander.citytowns, hotels and territories remain denied to the ru user.
• No 'hospital' trait exists in geonet; distance to a hospital needs an external source such as StatCan's Open Database of Healthcare Facilities.
• 0.0% of businesses here carry no coordinates, so a map or address lookup would miss them. Postal codes reach 99.9%.