What Lyon, France taught us about the hidden cost of stale location data

Dataplor analyzed 6 tourist cities across Europe and the US, tracking how often businesses opened and closed over time. We focused on the categories that move fastest and matter most to visitors: restaurants, bars, and entertainment.

One city ran away from the pack. Between August 3 and October 6, 2026, Lyon had the highest share of records updated of any city in the study: 8.7%—well ahead of the other 5.

Put another way, Lyon’s business landscape shifted about 14 times a day, every day, for 64 days straight. Over that window, we flagged 222 places as permanently closed and added 685 places to the dataset. That churn wasn’t concentrated in one trendy neighborhood; it touched all 9 arrondissements.

In Lyon, we list 1,383 categories across about 23,000 active and open places. The 685 additions and 222 closures highlighted on the map are spread across 295 unique place categories. However, the spatial distribution is more tightly packed in certain areas. The 2e and 3e, Lyon’s two biggest arrondissements by number of places, saw the most change by count, with 46 places flagged closed in the 2e on the Presqu’île and 165 added in the 3e around Part-Dieu.

Where change happens fastest

When we looked at which kinds of businesses close most often, the pattern was clear. In France, restaurants and bars are flagged closed roughly 3.5 to 5 times as often as accommodations or attractions.

Lyon is a food destination. Its tourist office counts 9.56 million overnight stays across the metropolitan area in recent years and 17 Michelin-starred restaurants. Over the 64 days, we found 70 restaurants and cafés and 9 bars flipped to permanently closed, against 13 attractions and recreation and 7 accommodation businesses. A full year of our French data shows the same order in the chart below.

The categories consumers interact with most frequently are the ones most likely to be wrong in a stale dataset. If you’re building a product around where people eat, drink, and spend their evenings, you’re building on the most volatile places in the city.

The visibility gap no one sees

Here’s where the cost of stale data becomes concrete.

Many teams start with open-source location data. It’s free, it’s broad, and it often looks complete. So we compared it against what we saw on the ground in Lyon.

In its September release, an alternative open-source dataset had no open-or-closed status at all for 99.6% of its Lyon listings. Not “closed,” not “open.” Simply unknown. A month later, a single bulk load from a business register filled in status for about 7 in 10 of those listings in one shot.

Think about what that means for anyone relying on it. For that September window, a product built on this data had essentially no way to tell a thriving restaurant from a shuttered one. And because the gap arrives in lumps rather than signaling itself, teams often don’t realize how long they’ve been flying blind.

Why freshness has to be continuous

Lyon makes a simple argument: in a city that changes 14 times a day, a dataset refreshed in occasional bulk loads is always behind, and you can’t know by how much.

The answer isn’t more data. It’s data that tracks change as it happens and tells you, record by record, whether a place is open right now. That’s what we built Dataplor to do. Our coverage of Lyon reflected those 222 closures and 685 additions across the 64 days as soon as they were detected, not after a registry caught up.

The difference shows up in a few practical ways. Fewer failed experiences for end users. Market sizing and territory planning based on businesses that actually exist. Risk and compliance decisions made on current status rather than last quarter’s snapshot. And, most importantly, confidence that the map you’re looking at is the city as it is today.

The question worth asking

Lyon isn’t an outlier because something strange is happening there. It’s an outlier because it’s a vibrant, fast-moving city, and fast-moving cities are exactly where businesses want to be. The same dynamics play out, at different speeds, in every market you operate in.

So here’s the question we’d put to anyone relying on location data: if 8.7% of the places in your most important city changed status in the last 2 months, would your data know? And if not, what is that gap costing you right now?

If you’d like to see how your markets compare, we’d be glad to show you what we’re seeing on the ground.

Sources

1. Lyon Tourist Office, “Tourism key figures,” last updated 2 June 2026. presse.lyon-france.com
2. Overture Maps Foundation, Places schema reference (operating_status) and Places guide. Schema, guide. Lyon rows from releases 2026-09-23.0 and 2026-10-21.0.
3. Map: rivers and boundaries from Overture Maps Foundation release 2026-10-21.0, which includes data © OpenStreetMap contributors (ODbL). Places: Dataplor.