Market research · 2026
LA Restaurant Market
Would a small robot-waiter café survive in Los Angeles — and if it works, can it scale into a chain?
Python · Pandas · Seaborn · EDA
The question
Investors wanted to open a café run by robot waiters — a high-capex concept that only pays back if the format can eventually become a chain. Before signing a lease, two things needed numbers: what the Los Angeles market actually looks like, and whether the chain route is plausible for a café at all.
The data
A registry of 9,651 food establishments in LA: name, address, whether it belongs to a chain, establishment type, and seat count. The base needed real cleaning before any proportion could be trusted — three rows with a missing chain flag were dropped, and 56 near-duplicate pairs (same address, same establishment under inconsistent spelling: "23RD ST CAFE" vs "23RD ST. CAFE") were resolved down to one row each. The working base closed at 9,592 establishments.
Name normalisation went one step further: 8,640 distinct raw names collapse into 7,773 brands once store-number suffixes and spelling variants are folded — Starbucks alone accounts for 106 raw spellings. Chain analysis was done on brands, not raw strings.
What the market says
Restaurants dominate: 75.2% of everything that serves food in LA. Cafés are a small slice — 4.4%, 426 establishments — which cuts both ways: less direct competition, but also a sign the format doesn't naturally thrive at scale there.
The chain question has a sharper answer. 62.1% of establishments are independent, but the mix inverts by type: every single bakery in the base belongs to a chain (100%), cafés are 60.6% chain, fast food 56.7% — while full restaurants (31.4%) and bars (26.4%) stay overwhelmingly independent. Cafés, in other words, are already a chain-friendly format in this market.
On size: the median establishment seats 27, with the middle half of the market between 14 and 46 seats. Chains do not run big rooms — they multiply small ones.
The recommendation
Open small and plan to replicate. A café of 25–35 seats matches both the market's median and the observed chain pattern: many modest venues rather than one large hall. The robot-waiter novelty carries the launch; the chain economics only work if the room stays small enough for the format to be stamped out repeatedly — which is exactly how the incumbent café chains in the registry behave.
Where the numbers live
Every figure on this page is computed in the sprint notebook, from the raw registry to the recommendation: Sprint10-mercado-de-restaurantes-LA.