Into the Night: Exploring the 24-hour Rhythms of the 15-Minute City

Into the Night: Exploring the 24-hour Rhythms of the 15-Minute City

This research examines the underexplored needs of nighttime communities in accessing urban amenities by foot. Using GIS-based spatiotemporal analysis, a customised version of Transform Transport’s 15-Minute City Score Toolkit, and Google Places API data, the study maps time-variant walkability in Milan. Findings reveal significant spatial and temporal gaps in nighttime accessibility and hourly calculations of the 15min City Score. Focusing on changes in levels of access between day and night, the research proposes a time-variant model of the 15-minute city framework based on actual service hours of different urban functions.

Introduction

Urban spaces are typically designed with a daytime bias. Services, amenities, and infrastructure are optimized for working hours, leaving nighttime often overlooked. From nightshift workers to cultural hubs, nightlife is vital to the social and economic fabric of cities. The night city significantly influences urban life, providing economic, cultural, and social opportunities, yet remains understudied in areas like accessibility and nighttime service provision. Research on the topic often focuses exclusively on transport services, issues of urban governance, or safety issues (GLA Economics, 2018; Seijas et al., 2024; Hadfield, 2015; Long & Ferenchak, 2021; McArthur et al., 2019).

This study explores Milan’s day-to-night service availability, emphasizing its impact on regular nighttime communities. These include essential workers, service industry staff, and creatives with predominantly nocturnal lifestyles (Brabazon, 2007). Despite diverse nighttime users, policies and economies often prioritize recreational nightlife consumers over workers’ needs (Acuto et al., 2021; Smeds et al., 2020; Furno et al., 2017). Night workers face challenges like limited late-night mobility, access to essential goods, and flexible childcare, crucial for a liveable nighttime city (Brusco et al., 2024; Kolioulis et al., 2021; Publica, 2020).

This research adopts a time-sensitive accessibility approach following the 15-minute city framework (Moreno et al., 2021; Teixeira et al., 2024; Olivari et al., 2023, integrating spatial and temporal dimensions of change. Unlike static proximity models, this research adopts a novel approach to the 15-minute city concept by factoring in temporal dynamics. This focus reveals that walkability and accessibility are not static but shift dramatically across different times of the day—a factor that holds profound implications for equity in urban planning.

Building on previous 15-minute City applications centred on physical proximity to urban functions, this study applies a time-variant approach to the urban accessibility narrative. By adopting this spatiotemporal approach, the study not only challenges static urban planning assumptions but also uncovers new insights into how the 15-minute city model could evolve to meet the needs of a 24-hour urban population.

Methodology: A Temporal Lens on Accessibility

The premise of the research is simple: amenities are only accessible when they are operational. The aim of this study is to tap into available information about the actual opening hours of diverse destinations in the city to highlight the relationship between walkability, accessibility to urban functions and real access conditions.

To highlight these differences, the research utilizes the 15-Minute City Score Toolkit, an open-source Python-based tool by Transform Transport designed to assess urban walkability and multiscale accessibility to services. It automates the calculation of the 15-Minute City Score, which focuses on access to amenities as the core metric, making it the appropriate tool to explore this topic (Albashir et al., 2024a; 2024b). For more information on the design and development of the 15-Minute City Score Toolkit, visit this article.

Key Tools and Data Sources

  • 15-Minute City Score Toolkit: An open-source Python-based tool designed to measure urban accessibility. The researchers extended its functionality to account for hourly variations in service availability, enabling a dynamic analysis of accessibility scores across the day.
  • Google Places API: Used to gather data on Points of Interest (POIs), including their types, locations, and operational hours. This allowed the team to map when and where services were available within walking distances.
  • H3 Hexagonal Grids: A hierarchical spatial framework that divides Milan into 1741 uniform hexagons. Each hexagon represented a localized analysis unit, facilitating the study of service availability at a granular level.

Google Places API data was selected for its detailed opening hour information and comparable completeness to OSM data (Balducci, 2021; Touya et al., 2017). With over 46,000 POIs retrieved, the data was cleaned and processed to create a final structured dataset. Supplementary GTFS data was used for public transport POIs, ensuring accurate schedules for June 2024. The process involved:

  • Zoning Algorithm: Using H3 indexing, Milan was divided into hexagonal grids. Queries were refined iteratively to fit the API’s 20-result limit.
  • Data Retrieval: HTTP POST requests retrieved POI details, and JSON responses were parsed to extract unique entries.
  • Data Cleaning: Duplicates were removed, and key attributes such as coordinates, types, and opening hours were standardized into a CSV format.

From Data to Insights: A Toolkit Tailored for Time-Variant Analysis

Adapting the toolkit for this purpose required a series of methodological innovations. The original toolkit was designed around OpenStreetMap (OSM) data, which excels in mapping physical infrastructure but lacks information on when services are actually available. To address this, the researchers turned to the Google Places API, a more dynamic data source offering detailed information on service locations, types, and opening hours. The API yielded a dataset of over 46,000 points of interest (POIs) across Milan including opening hour information, representing around 58% of all POIs available in the city of Milan.

Given the breadth of data from the Google Places API, careful structuring was essential. The POIs were organized into two hierarchical levels for clarity and usability. Place types (T_L2) were classified into broad macro-categories (T_L1) like “Food and Drink” or “Health” to allow for large-scale analysis. Grouping amenities in relation to their relation to daily urban activities is common practice in 15-minute City research and practical applications (Moreno et al., 2021; Abdelfattah et al., 2022). This categorization helped reconcile differences between Google’s taxonomy and the toolkit’s original classifications, ensuring that the study’s findings were both consistent and meaningful.

POIs were filtered based on relevance, specificity, and consistency. These were revised to resolve discrepancies in multi-label classifications, eliminate overestimated or mislabeled entries, and ensure data alignment with ground truth sources (e.g., municipal datasets). The final dataset included a total of 160 place types (T_L2) across nine macro-categories (T_L1): Services, Food & Drink, Shopping, Health, Education, Sports, Culture & Entertainment, Open Leisure, and Mobility.

The study divided Milan into 1,741 hexagonal grid cells using the H3 Hexagonal Hierarchical Spatial Index, a system well-suited for spatial analysis. Each hexagon was assigned a 15-minute walking radius based on a pedestrian network, with an average walking speed of 4.5 km/h. This grid-based approach allowed for high-resolution analysis of service accessibility across the city.

A key feature of the toolkit is its exponential distance decay function, which mimics real pedestrian behavior by weighting closer services more heavily than distant ones. However, in the original architecture of the toolkit, only the nearest function under each class macro-category is considered, leading to a potential overrepresentation of nighttime accessibility given the reportedly inflated number of places in the dataset listed as open 24 hours (5% of all POIs in the dataset). For that reason, a second approach was adopted in which all places within a service class are included in the analysis, thereby diluting the impact of 24-hour open places and highlighting the role of service density in inter-district comparisons.

To enable a time-variant analysis of Milan, four key modifications were made to the 15-min Score Tool:

  1. Data Source: Proprietary Google Places API data replaced OSM to incorporate service opening hours.
  2. Taxonomy: Google Places’ classifications were restructured to align with the toolkit’s macro-categories. (See Table 1 for reclassification).
  3. Methodology: The scoring method was refined to include all reachable amenities within 15 minutes, enabling weighted scores to exceed 1, unlike the original design.
  4. Zoning: H3 hexagonal grids (resolution 9) were used for spatial analysis, dividing Milan into 1,741 hexagons with precise isochrones for walkability assessment.

Key Findings: Gaps in Day-Night Accessibility

The inclusion of opening hours offered fresh insights into Milan’s urban dynamics. For example, “Food and Drink” services had the highest proportion of POIs with complete opening hour data (86%), allowing for detailed analysis of their contribution to morning and evening accessibility. In contrast, categories like “Open Leisure” and “Mobility” lagged behind, with only 42% and 47% (excluding public transport) of POIs, respectively, providing reliable opening hour data, highlighting potential data gaps in the Google Places dataset for non-commercial destinations.

Figure 1 Hourly distribution of open places by macro-category on a typical weekday by total counts, and individually on a normalized scale

Weekday Service Patterns

  • The highest number of open places is at 12:00, while the lowest is at 03:00, where only 6% of midday places remain open. Public-facing categories such as Open Leisure (81%) and Mobility (23%) maintain relatively higher nighttime availability compared to private indoor categories like Shopping and Food & Drink.
  • Most categories exhibit double-peak activity, with highs in the morning and late afternoon, and declines during lunchtime and evening hours. Exceptions include Food & Drink, which peaks during midday, and Mobility, which remains relatively consistent.

Nighttime Access and Urban Equity

  • At midnight, the availability of services for nighttime users drops significantly, with access to Food & Drink, Sports, and Mobility services at only 14–23% of midday levels. Essential services such as Education, Health, and Shopping fall to just 3–4%.
  • This disproportionate availability underscores challenges for night workers and leisure communities in clustering essential activities around their atypical urban use patterns dominating in the nighttime hours.

Weekend Variations

  • On Saturdays, service availability reaches 80% of typical Wednesday levels, while Sundays drop to 40%. Cultural and entertainment activities extend further into the night on weekends, while Health services shift toward morning availability.
  • Sundays exhibit a flatter distribution, with potential for some categories showing anomalous nighttime shares due to overreported 24-hour services.
Figure 2 Heatmap showing relative hourly distribution of open places by macro-category on a typical weekday (Wednesday) and weekend (Saturday, Sunday) in Milan

24-Hour Services

  • Only 5% of places are listed as open 24 hours, dominated by public-oriented amenities like electric vehicle charging stations, parks, and gas stations. Misreported 24-hour services are predicted based on the voluntary contribution of opening hours by businesses and private individuals and the commercial nature of Google Places.
  • These services represent over 40% of all open places during late-night hours, peaking at 90% at 03:00, but account for just 11% of daytime availability, presenting significant potential for the distortion of results.

Spatial Analysis with the 15-Minute City Score

  • Central areas maintain relatively high service accessibility throughout the day, while peripheral neighbourhoods experience more drastic drops at night. The largest gaps are seen in the southeast (e.g., Chiaravalle) and northeast areas.
  • The density-based analysis revealed sharper disparities in central areas, with accessibility falling by up to 95% in some neighbourhoods (e.g., Duomo and Porta Venezia) between 12:00 and 03:00, signalling sharper disparities in the breadth of options within the same urban functional macro category in these areas.
Figure 3 15min City Score at peak hours 12:00 and 03:00 and difference between the two hours, using (a) nearest-amenity method, and (b) density-based method.

Public Transport and Accessibility

  • Waiting times for public transport increase significantly at night. While most areas have 5–15 minute waiting times at noon, waiting times increase to 15–20 minutes in central areas, while over 18% of outer zones experience waits of up to 30 minutes, emphasizing the need for improved night transit options.
Figure 4 Average maximum waiting time for Public Transport by hexagon at 12:00 and 03:00.

Implications for Urban Equity: Insights from the Milan case study

These findings reveal the stark contrast in service availability between daytime and nighttime, highlighting challenges for equitable access to diverse urban functions in Milan. Improving nighttime accessibility, particularly in peripheral neighbourhoods and for critical services like Health and Public Transport, is crucial for fostering a city that meets the needs of all residents, including night workers and wider nocturnal communities.

Addressing these gaps requires a paradigm shift in how we think about the role of time in urban planning. Milan’s Territorial Timetable Plan (PTO) provides a foundation for aligning service schedules with asynchronous users’ needs, promoting equity and liveability for nighttime communities (Bonfiglioli, 1997; Comune di Milano, 2024). The Milan case study highlights the adaptability and potential of the 15-Minute City Score Toolkit. By integrating temporal data and refining its analytical framework, the toolkit proved capable of addressing complex questions about urban accessibility. It offers a scalable solution for cities worldwide, enabling researchers and policymakers to dynamically assess the spatiotemporal distribution of accessibility and the inclusivity outcomes for underrepresented urban groups.

A Call to Action: Towards New Urban Time Policies

The findings from Milan underscore the need for cities to embrace a more holistic approach to the 15-minute city. This includes:

  • Prioritizing Nighttime Services for Night Workers: Policymakers must expand night services beyond the needs of cultural consumers, developing equitable urban time policies that ensure a fair and balanced distribution of essential services to predominant night users.
  • Integrating Temporal Data into Planning: Urban accessibility tools and applications should incorporate time-sensitive data to ensure that planning decisions account for nighttime needs.
  • Engaging Nighttime Communities: Planners should collaborate with nocturnal communities, including night workers, businesses, and residents to better understand their unique needs and challenges.
  • Expanding the Definition of Essential Services: Services like childcare, pharmacies, and convenience stores should be considered critical for nighttime accessibility, to ensure the adequate functioning of night cities for asynchronous users.
  • Encouraging Localized Solutions: Neighbourhood-specific strategies could help mitigate spatio-temporal inequities, with tailored solutions to address the unique challenges of each area.

The results of this research have been published in the peer-reviewed, open-access journal Computers: Abdelfattah, L., Albashir, A., Ceccarelli, G., Gorrini, A., Messa, F., & Presicce, D. (2025). The Right to the Night City: Exploring the Temporal Variability of the 15-min City in Milan and Its Implications for Nocturnal Communities. Computers, 14(1), 22. https://doi.org/10.3390/computers14010022

References

Abdelfattah, L., Deponte, D., & Fossa, G. (2022). The 15-minute city: Interpreting the model to bring out urban resiliencies. Transportation Research Procedia, 60, 330–337. https://doi.org/10.1016/j.trpro.2021.12.043

Acuto, M., Seijas, A., McArthur, J., & Robin, E. (2021). Managing cities at night: A practitioner guide to the urban governance of the night-time economy. Bristol University Press. https://doi.org/10.51952/9781529218305

Albashir, A., Messa, F., Presicce, D., Pedrazzoli, A., & Gorrini, A. (2024a). 15min City Score Toolkit – Urban Walkability Analytics. https://doi.org/10.5281/ZENODO.14231533

Albashir, A., Messa, F., Presicce, D., Pedrazzoli, A., & Gorrini, A. (2024b). 15min City Score Toolkit—Notebook [Computer software]. Zenodo. https://doi.org/10.5281/ZENODO.14231427

Balducci, F. (2021). Is OpenStreetMap a good source of information for cultural statistics? The case of Italian museums. Environment and Planning B: Urban Analytics and City Science, 48(3), 503–520. https://doi.org/10.1177/2399808319876949

Bonfiglioli, S. (1997). Urban time policies in Italy: An overview of time-oriented research. Transfer: European Review of Labour and Research, 3(4), 700–722. https://doi.org/10.1177/102425899700300405

Brabazon, T. (2007). Into the night-time economy: Work, leisure, urbanity and the creative industries. Nebula, 4(2), 161–178.

Brusco, A., Bucci, G., Campea, S., Filottrano, A., Marracino, F., & Romualdi, G. (2024). Gli infortuni sul lavoro in orario notturno in Italia. Inail Consulenza Statistico Attuariale. https://www.inail.it/content/dam/inail-hub-site/documenti/catalogo-generale/2024/07/GliInfortuniSulLavoroInOrariNotturnoInItalia.pdf

Comune di Milano. (n.d.). I tempi della città: Il Piano Territoriale degli Orari. Retrieved 18 December 2024, from https://www.comune.milano.it/aree-tematiche/lavoro-e-formazione/i-tempi-della-citta-il-piano-territoriale-degli-orari-pto#:~:text=Il%20P.T.O.,alla%20loro%20armonizzazione%20e%20coordinamento

Furno, A., El Faouzi, N.-E., Fiore, M., & Stanica, R. (2017). Fusing GPS probe and mobile phone data for enhanced land-use detection. In Proceedings of the 2017 5th IEEE International Conference on Models and Technologies for Intelligent Transportation Systems (MT-ITS) (pp. 693–698). IEEE. https://doi.org/10.1109/MTITS.2017.8005601

GLA Economics. (2018). London at night: An evidence base for a 24-hour city. GLA Economics.

Hadfield, P. (2015). The night-time city: Four modes of exclusion: Reflections on the urban studies special collection. Urban Studies, 52(3), 606–616. https://doi.org/10.1177/0042098014552934

Kolioulis, A., Siravo, J., Apostolidis, P., Kummer-Buléon, C., Matheou, L., & Campani, C. (2021). Working nights: Municipal strategies for nocturnal workers. Autonomy. https://autonomy.work/wp-content/uploads/2021/12/Working_Nights-2.pdf

Long, B., & Ferenchak, N. N. (2021). Spatial equity analysis of nighttime pedestrian safety: Role of land use and alcohol establishments in Albuquerque, NM. Transportation Research Record: Journal of the Transportation Research Board, 2675(10), 622–634. https://doi.org/10.1177/03611981211030263

McArthur, J., Robin, E., & Smeds, E. (2019). Socio-spatial and temporal dimensions of transport equity for London’s night time economy. Transportation Research Part A: Policy and Practice, 121, 433–443. https://doi.org/10.1016/j.tra.2019.01.024

Moreno, C., Allam, Z., Chabaud, D., Gall, C., & Pratlong, F. (2021). Introducing the “15-minute city”: Sustainability, resilience and place identity in future post-pandemic cities. Smart Cities, 4(1), 93–111. https://doi.org/10.3390/smartcities4010006

Olivari, B., Cipriano, P., Napolitano, M., & Giovannini, L. (2023). Are Italian cities already 15-minute? Presenting the next proximity index: A novel and scalable way to measure it, based on open data. Journal of Urban Mobility, 4, 100057. https://doi.org/10.1016/j.urbmob.2023.100057

Publica. (2020). Developing a night time strategy. Part 1: Guidance on process. Greater London Authority. https://www.london.gov.uk/sites/default/files/210317_gla_1_night-time_strategies_part_1.pdf

Seijas, A., Barnett, J., & Salihudin, S. (2024, January 9). Rethinking 24-hour cities: Night-time strategies to address urban challenges and thrive. World Economic Forum. https://www.weforum.org/stories/2024/01/24-hour-cities-night-time-strategies-urban-challenges/

Smeds, E., Robin, E., & McArthur, J. (2020). Night-time mobilities and (in)justice in London: Constructing mobile subjects and the politics of difference in policy-making. Journal of Transport Geography, 82, 102569. https://doi.org/10.1016/j.jtrangeo.2019.102569

Teixeira, J. F., Silva, C., Seisenberger, S., Büttner, B., McCormick, B., Papa, E., & Cao, M. (2024). Classifying 15-minute cities: A review of worldwide practices. Transportation Research Part A: Policy and Practice, 189, 104234. https://doi.org/10.1016/j.tra.2024.104234

Touya, G., Antoniou, V., Olteanu-Raimond, A.-M., & Van Damme, M.-D. (2017). Assessing Crowdsourced POI Quality: Combining Methods Based on Reference Data, History, and Spatial Relations. ISPRS International Journal of Geo-Information, 6(3), 80. https://doi.org/10.3390/ijgi6030080