🔬 physics

The benefits and biases of seeing the word's cities through marathons

By analyzing 311 marathon routes across five continents, this study reveals that marathon courses are not representative of cities but instead function as intentionally selective urban portraits that disproportionately feature museums and luxury brands compared to other areas or alternative routes.

Original authors: Andrew Renninger

Published 2026-03-24

📖 5 min read🧠 Deep dive

CC0 1.0

Original authors: Andrew Renninger

Original paper dedicated to the public domain under CC0 1.0 (http://creativecommons.org/publicdomain/zero/1.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine you are visiting a new city for the first time. You decide to run a marathon there. You spend four hours running, sweating, and looking at the world as you pass by. By the time you cross the finish line, you feel like you "know" that city. You have a mental map of its streets, its sights, and its vibe.

But here is the twist: That mental map is a carefully curated highlight reel, not the whole movie.

This paper by Andrew Renninger is like a detective story that investigates exactly what marathon runners see when they run through 311 cities around the world. The author asks a simple question: If a marathon route is a camera lens, what kind of city is it photographing?

Here is the breakdown of the findings, using some everyday analogies:

1. The "Greatest Hits" Album vs. The Full Discography

Think of a city like a music band. A city has deep, complex tracks (schools, grocery stores, auto repair shops, daycares) and flashy, polished hits (museums, luxury boutiques, famous monuments).

The study found that marathon routes are like a radio station that only plays the Greatest Hits.

The Bias: If you run a marathon, you are 15 times more likely to run past a museum than a random person walking around the city would be. You are 8.5 times more likely to see a luxury brand store.
The Missing Tracks: You are much less likely to see the "everyday" stuff. Schools, hardware stores, and regular grocery stores are often left out of the frame. It's as if the marathon route is a tour guide who only takes you to the fancy hotels and the famous statues, skipping the neighborhood where people actually live, work, and buy milk.

2. The "Instagram Filter" Effect

The author argues that marathons act like a heavy-handed Instagram filter. They make the city look more glamorous, historic, and tourist-friendly than it really is.

The Luxury Lens: In cities like Florence or Milan, the route is practically a runway for high-end fashion. The study found that in some Italian cities, there are dozens of luxury boutiques right on the running path, but almost none just a few blocks away.
The Apple Store Clue: The researchers even looked at Apple Stores. Because Apple deliberately places its stores in the most "iconic" and central spots, they act as a perfect test. The study found that marathon runners are significantly closer to an Apple Store than the average city dweller is. The route is drawn to the "cool" spots.

3. It's Not Just Because the City is Busy

You might think, "Well, marathons just run through the busy downtown because that's where the roads are wide and flat."

The author tested this by creating a "What If" scenario (a null model). Imagine you have a marathon that starts at the Eiffel Tower and ends at the Louvre. The computer generated 50 different other possible routes that also started and ended there but took different paths through the city.

The Result: Even when the computer picked a different path, the actual marathon route chosen by humans was still more packed with museums and tourist traps than the random alternatives.

The Analogy: It's like planning a road trip from New York to Boston. You could take a backroad that goes through quiet farms and small towns. But the marathon planners chose the highway that goes through the most famous landmarks. They didn't just pick the easiest path; they picked the most scenic path.

4. The Loop vs. The Straight Line

The study also noticed something interesting about the shape of the race:

Loop Races (Start and finish at the same spot): These are like a tour bus that can drive in circles. They have the most freedom to detour and show off the "best" parts of the city. These races showed the strongest bias toward tourist spots.
Point-to-Point Races (Start and finish far apart): These are like a train that has to get from A to B. They have less room to wiggle around to see the fancy stuff, so they are slightly less biased, but they still favor the "tourist" view over the "local" view.

The Big Takeaway

Marathons are wonderful, exciting, and great for tourism. But they are not neutral.

When you run a marathon, you aren't seeing a "fair cross-section" of the city. You are seeing a curated exhibition. The organizers are essentially saying, "We want you to see the history, the art, the wealth, and the beauty of this place."

They are leaving out the schools, the laundromats, the local hardware stores, and the everyday neighborhoods where the city's actual life happens. If you only know a city through its marathon, you are seeing the city's "glossy brochure" version, not its messy, real, everyday self.

In short: Marathons are the ultimate "highlight reel" of a city. They show you the best parts, but they hide the rest.

1. Problem Statement

While marathon running has surged in popularity post-pandemic, serving as a primary mode of urban exploration for tourists and locals alike, there is a lack of empirical understanding regarding how representative marathon routes are of a city's actual urban fabric.

The Core Question: Do marathon routes offer a neutral cross-section of urban life, or do they systematically privilege specific districts, amenities, and demographics?
The Implication: If routes are biased, they shape not only what runners remember but also the "destination image" projected to visitors, potentially reinforcing stereotypes of cities as centers of tourism and luxury while obscuring everyday residential and service-oriented areas.

2. Methodology

The study employs a quantitative spatial analysis comparing observed marathon routes against both the surrounding city and statistically generated "null" routes.

Dataset:
- Sample: 311 marathon routes across five continents (concentrated in Europe and North America).
- Data Sources: GPX files from Go&Race; street networks and landmarks from OpenStreetMap (via OSMnx); amenities, brands, and points of interest (POIs) from Foursquare Open Source Places; built environment data from GHSL and Sentinel-2.
Spatial Definitions:
- On-Route: A 200-meter corridor surrounding the race path.
- Off-Route (Surrounding City): A 5-kilometer buffer around the route, excluding the 200-meter corridor. This represents the "plausible alternative" urban fabric the race could have traversed.
Analytical Metrics:
- Density Ratio ( $R_{dens}$ ): Compares the density of a specific amenity category on-route vs. off-route.
  $R_{dens} = \frac{N_{on}/A_{on}}{N_{off}/A_{off}}$
  To isolate selective bias from general centrality, the authors normalize this ratio by the geometric mean of all amenity ratios within the same city.
- Proximity Ratio ( $R_{prox}$ ): Compares the mean nearest distance from route streets to a specific amenity vs. the mean distance from all city streets.
- Principal Component Analysis (PCA): Used to reduce dimensionality and identify underlying patterns (e.g., distinguishing "centrality" from "leisure/tourism" bias).
Null Model (Perturbation Analysis):
- To determine if biases are merely a function of starting/finishing in city centers, the authors generated 50 alternative routes for each marathon.
- Process: Fixed start and finish points; intermediate nodes were perturbed using biased random walkers on the street network.
- Constraints: Alternative routes were rejected if they deviated by more than 10% from the official 42.2 km distance.
- Comparison: Observed routes were compared against the distribution of these null routes using Z-scores.

3. Key Results

A. General Urban Bias

Marathon routes consistently traverse denser, taller, and more urbanized areas than the surrounding city:

Population density and building height are 1.25 times higher on-route.
Retail and dining amenities are 8.5 times denser per square kilometer on-route.

B. Systematic Categorial Bias

Even after controlling for general density, specific categories are disproportionately overrepresented:

Cultural & Tourist Landmarks: Museums are 15.7 times denser on-route; memorials, attractions, statues, and galleries show similar massive overrepresentation (11–13x).
Leisure vs. Everyday Services:
- Overrepresented: Museums (2.5x above city average), galleries, hotels, bars, and restaurants.
- Underrepresented: Schools, daycares, hardware stores, auto repair shops, and supermarkets.
- Composite Skew: The median route is 2.7 times more skewed toward leisure/tourism than toward everyday services.
Commercial Landscape:
- Chain Stores: Surprisingly, chain stores are less common on-route (median 4.5%) than off-route (5.9%).
- Luxury Brands: A sharp reversal occurs for luxury. The median city has 25.8 times more luxury brands within 200m of the route than the rest of the city. (e.g., Florence has 44 luxury boutiques on-route vs. 3 off-route).
- Apple Stores: Routes pass significantly closer to Apple Stores (1.68x closer than average city streets), reflecting the brand's strategy of placing stores in historic/central "destinations."

C. Robustness via Null Model

The bias persists even when comparing observed routes to plausible alternatives with the same start/finish:

Landmarks: 82% of cities have higher landmark ratios on the observed route than the mean null route. The median observed course is 1.79 times richer in landmarks than the null mean.
Amenities: 68% of cities show higher amenity ratios.
Statistical Significance: 53% of races exceed the null model at $p < 0.05$ for landmarks.
Route Type: Loop races (start/finish adjacent) show higher bias than point-to-point races, as loops allow more freedom to detour through preferred districts.

4. Key Contributions

Quantification of "Curated" Urbanism: The paper provides the first large-scale empirical evidence that marathon routes are not accidental but are "intentionally selective urban portraits." They function as curated exhibitions that decide which streets represent the city.
Methodological Framework: Introduces a robust null-model approach (perturbing routes while fixing anchors) to disentangle the effects of centrality (running through the middle of a city) from selection bias (choosing specific scenic or prestigious streets within that center).
Differentiation of Bias Types: Distinguishes between "centrality bias" (all amenities are denser) and "functional bias" (tourism/leisure amenities are disproportionately denser compared to everyday services).
Commercial Insight: Reveals a nuanced commercial bias where mass-market chains are avoided in favor of high-prestige luxury brands and "destination" retailers (like Apple).

5. Significance

For Urban Planning & Tourism: The findings suggest that marathon tourism reinforces a "postcard" version of a city, potentially distorting visitor perceptions of urban life. It highlights how event zones can concentrate visitors in narrow, high-prestige corridors, excluding the "ordinary" city.
For Social Geography: The study underscores that mobility shapes exposure and inference. Runners (and the cities they represent) are systematically exposed to the "symbolic city" (monuments, luxury) while remaining blind to the "functional city" (schools, repair shops, local grocery).
For Future Research: Establishes a baseline for analyzing other forms of event-based mobility (e.g., cycling tours, walking festivals) and their role in shaping destination images and social attitudes.

Conclusion: Marathons are not neutral passages through a city; they are highly curated experiences that prioritize scenic, historic, and luxury elements, effectively filtering out the mundane realities of daily urban life.