Data-Driven Analysis of Bicycle Lane Safety in Mexican Cities:

Towards a Real-Time Route Recommendation System for Cyclists

Carlos Alberto Domínguez-Báez

1,2 a

, Ricardo Mendoza-González

and Huizilopoztli Luna-García

Tecnológico Nacional de México / IT Aguascalientes, Av. Adolfo López Mateos, 1801 Ote., Aguascalientes, Mexico

Universidad Autónoma de Zacatecas, Ramón López Velarde 801, Zacatecas, Mexico

Keywords: Data-Driven Analysis, Safe Bicycle Lanes, Urban Eco-Mobility.

Abstract: This study initiated a project to identify urban cycling routes with a focus on cyclist safety in Mexican cities.

A Data-Driven Analysis approach was implemented to map the riskiest and safest cycling routes by analysing

traffic accident data from national, state, and local datasets. The accident hotspots were visually integrated

into the urban map of Guadalajara city (Jalisco, Mexico), to identify high-risk zones for cyclists. The

integration of diverse data sources and geospatial analysis allowed for an accurate characterization of accident

patterns, providing a clear identification of critical areas. Key results from this initial stage of the project

included an accurate risk-zones identification, a replicable methodology for data integration, and a first

approach to developing algorithms for cyclist accident analysis. These preliminary findings hold promise for

enhancing urban cycling safety and supporting urban eco-mobility strategies in Mexican cities. Additionally,

the results served as a foundation for future exploration of machine learning techniques to refine data

processing and develop a real-time safe bicycle lane recommender prototype aimed at guiding cyclists toward

safer alternatives.

1 INTRODUCTION

Effective transportation system planning must

address not only the mobility needs of individuals but

also ensure safety and sustainability, promoting eco-

friendly strategies. According to the World

Commission on Environment and Development,

sustainable development involves meeting the needs

of the present without compromising the ability of

future generations to meet their own needs. In this

context, transportation systems play a crucial role in

sustainable development by facilitating access to

economic and social opportunities, which is essential

for balanced growth across economic, social, and

environmental spheres (Visser, 2017).

One of the major challenges in creating

sustainable transportation networks is overcoming

natural barriers and reducing environmental impact

(Mahfouz, et al., 2023). Transportation infrastructure

must evolve to ensure sustainable mobility, not just

https://orcid.org/0000-0001-8820-5518

https://orcid.org/0000-0002-8934-8067

https://orcid.org/0000-0001-5714-7482

through the construction of adequate road networks,

but also by integrating solutions that minimize

environmental harm (Bahmankhah & Coelho, 2017).

Urban transport sustainability is closely linked to the

implementation of innovative systems that foster

public trust and promote less polluting transport

modes, such as cycling (Bahmankhah & Coelho,

2017; Černá, et al., 2014).

Cycling has emerged as a prominent option within

urban eco-mobility strategies (Ogryzek, 2020).

Although in many developing countries, bicycles are

primarily used for recreational purposes, their

potential as a daily transportation mode is significant

(Heesch & Sahlqvist, 2013). Promoting cycling can

drastically reduce CO2 emissions, alleviate traffic

congestion, and improve air quality (Nasir, 2024;

Batool, et al. 2024). Bicycles, as an eco-friendly

mode of transport, play a key role in reducing vehicle

emissions, minimizing congestion, and lowering

transportation costs, in addition to benefiting the

Domínguez-Báez, C. A., Mendoza-González, R. and Luna-García, H.

Data-Driven Analysis of Bicycle Lane Safety in Mexican Cities: Towards a Real-Time Route Recommendation System for Cyclists.

DOI: 10.5220/0013365100003941

Paper published under CC license (CC BY-NC-ND 4.0)

In Proceedings of the 11th International Conference on Vehicle Technology and Intelligent Transport Systems (VEHITS 2025), pages 543-548

ISBN: 978-989-758-745-0; ISSN: 2184-495X

543

physical and mental health of users (Gulati, 2024).

Numerous studies have shown that policies

promoting cycling significantly improve urban

quality of life. As a result, many governments

worldwide are implementing initiatives to encourage

its use (Kosmidis & Müller-Eie, 2024).

Despite the numerous benefits of cycling, its

integration into cities must account for the safety of

cyclists and pedestrians, which is compromised in

cities where cycling infrastructure is inadequate or

poorly located (Liu, et al., 2024). The development of

safe and dedicated cycling infrastructure, such as

exclusive bike lanes, is critical to encouraging cycling

(Khademi, et al., 2024). While some studies suggest

that increased infrastructure availability can boost

cycling trips, it is also emphasized that simply

improving infrastructure is not enough; these spaces

must be safe, particularly for cyclists and pedestrians

and other aspects need to be addressed too, such as

perceptions of insecurity, and cultural barriers (Al-

Ansari & Al-Khafaji, 2024; Khademi, et al., 2024).

This necessitates planning that includes physical

separation from motorized traffic, speed reductions,

and educational campaigns that promote road safety

from an early age (Toski, et al., 2024).

Moreover, social attitudes and norms play a

pivotal role in the adoption of sustainable transport

modes (Kočková, et al., 2024). People living in bike-

friendly environments, such as cities with adequate

infrastructure, are more likely to use bicycles

frequently than those in areas with fewer facilities

(Khademi, et al., 2024; Useche, et al. 2024).

Promoting favourable attitudes toward cycling

through public policies and community support is

essential to shifting transportation habits (Khademi,

et al., 2024; Kočková, et al., 2024).

In several countries such as Mexico, strategies

are being successfully implemented in favour of the

use of bicycles as a means of transportation, such as:

EcoBici in Mexico City (Peralta, 2016), and MiBici

(https://www.mibici.net/), in Guadalajara; however,

these alternatives tend to be visualized only in a few

large cities. In most of the territory the needs of

infrastructure, social awareness, and safety are still

very evident (Lagunas-Millan, 2018). Investments in

cycling infrastructure often fail to meet the needs of

regular cyclists, who are mostly from low-income

backgrounds (Lagunas-Millan, 2018; Peralta, 2016).

Additionally, the lack of connectivity and poorly

planned cycling routes expose cyclists to safety risks,

as they are forced to share roads with motorized

traffic without proper lanes (Lagunas-Millan, 2018;

Peralta, 2016). This highlights the need for planning

and forecasting tools that integrate built environment

characteristics to prioritize infrastructure investments

in high-need areas (Etminani-Ghasrodashti, 2018).

In this context, optimizing cycling routes and

infrastructure is crucial to maximizing the benefits of

cycling mobility (Komarica, et al. 2024). Advanced

technologies such as Artificial Intelligence (AI) and

Machine Learning (ML) enable the analysis of data

on road safety, traffic flow, and emissions to design

more efficient solutions (Komarica, et al. 2024;

Younus, et al. 2025). AI can help identify optimal

routes for cyclists by considering variables such as

safety, travel time, emissions, and user preferences

(Younus, et al. 2025). Integrating multi-objective

optimization enables balancing traffic performance,

safety, and emissions, thereby promoting the use of

more sustainable transport modes like cycling

(Avina-Bravo, et al. 2022; Koska et al. 2021).

In summary, advancing toward more eco-friendly

and safe urban mobility requires integrating

innovative technologies that optimize infrastructure

and urban planning. These solutions must ensure the

safety of cyclists and pedestrians, maximize

transportation system efficiency, and reduce harmful

emissions. Cycling, as an eco-friendly transport

mode, plays a vital role in this process, and its

effective integration into cities depends on policies

and technologies that consider local characteristics

and user needs.

To address this gap, we have initiated a project

aimed at identifying urban cycling routes with an

emphasis on cyclist safety in Mexican cities. From an

initial literature review (see Section 2 for

methodology details), we identified four key papers

that are directly relevant to this project. A summary

of these related works is provided below, ordered by

relevance, to establish the context and highlight their

contribution to the current research.

1. Seoudi et al. (2023) proposed a multi-criteria

route planning strategy that optimizes the

comfort, health, and safety of cyclists. Their

system operates independently of specific bike

lane information or traffic regulations,

incorporating real-time weather data to

improve the optimization of urban cycling

routes.

2. Ferreira and Costa (2024) developed an

innovative low-cost integrated system to

improve cycling safety in urban environments.

The system assesses proximity to emergency

services and uses GPS coordinates to determine

dynamic levels of risk for cyclists, generating

real-time alerts when crossing high-risk areas.

3. The research of (Pindarwati & Wijayanto,

2019) describes an integrated web-based

VEHITS 2025 - 11th International Conference on Vehicle Technology and Intelligent Transport Systems

544

system for personalized navigation that uses

crime hotspot data from local agencies, social

media, and user reports. The system offers

rerouting options and a crime risks zones map,

aiming to recommend safer routes to users.

4. Chavez et al. (2019) presented the Safe

Commuting System (SCS), a solution designed

to improve urban commuting safety through

crowdsourced mobile device data. The system

implements real-time user alerts for safety

incidents and provides alternative routing

options by categorizing data into three main

domains: criminal activity, perceived danger,

and suspicious behaviour.

These studies serve as a foundational basis for the

proposed approach, offering valuable insights and

guiding the development of insights to evolve the

proposal. In this case, for this initial stage of the

project, we mapped the riskiest and safest cycling

routes by analysing traffic accident databases

available at national, state, and city levels. This

analysis was then visually represented by overlaying

accident hotspots onto the urban map of Guadalajara,

Jalisco, Mexico. The results from this stage are

expected to serve as the foundation for developing a

real-time urban cycling route recommender system.

This system will guide cyclists by suggesting safer

alternative routes and providing safety-related

information, thus contributing to improved urban

mobility and eco-friendly transportation solutions.

The rest of the article is organized as follows:

Section 2 describes the methodology employed to

carry out this stage of the project. Section 3 details the

development of the first stage and presents the results

obtained so far. Finally, Section 4 offers the

conclusions of this stage and outlines potential

directions for future work.

2 METHODOLOGY

A literature review was conducted using the Semantic

Scholar tool (https://www.semanticscholar.org). This

review led to the identification of a final pool of 22

relevant sources (out of an initial pool of 669),

comprising 12 peer-reviewed research articles, eight

conference papers from international forums, one

international project report, and one article from a

blog. The selected sources were filtered based on their

relevance to the topic, considering citation counts and

recency, with a focus on publications from 2017 to

2024.

The analysis of the selected literature provided

insights into the general strategy for constructing the

accident analysis database (BD) and highlighted

various technological strategies that could be

implemented in subsequent project phases.

As part of the methodology, an initial database

was created using official traffic accident and

vehicular flow records from the following sources:

1. National Institute of Statistics and Geography

(INEGI, Mexico): The INEGI database serves

as a primary source of national traffic accident

data in Mexico. This dataset includes 43

distinct fields, including: a) Geospatial

information: accident latitude and longitude;

temporal data: year, month, day, hour, and

minute; b) Accident characteristics: type,

cause, and road conditions; c) Vehicle

information: types (including bicycles) and

numbers of vehicles involved; d) Victim data:

number of injuries and fatalities by user type;

e) Contextual information: type of roadway,

urban/suburban conditions. The detailed

structure of the database allowed for an in-

depth analysis of the specific circumstances

surrounding each incident, facilitating the

identification of contributing factors and

patterns.

2. Institute of Statistical and Geographical

Information of Jalisco (IIEGJ, Mexico): This

regional database complements the national

dataset (BD1) with local context-specific

details. It includes 20 fields, focusing on: a)

Unique incident identifiers; b) Detailed

temporal data; c) Precise location with cross-

referenced street information; d) Detailed

accident typology, such as cyclist involvement;

e) Demographic characteristics of the involved

individuals; f) Specific consequences of the

incidents. This dataset is particularly valuable

due to its regional focus, providing contextual

information specific to the metropolitan area of

Guadalajara (Jalisco, Mexico).

3. “GDL en Bici Program”, Jalisco State

Government (Mexico): This database is part of

a comprehensive traffic infrastructure dataset,

containing geospatial data critical to the

existing network of bicycle lanes. It operates

within a GeoJSON framework, including a)

Precise geometry of bicycle lanes; b)

Infrastructure attributes such as the type of

segregation; c) Temporal information on

construction and modifications; d) Design

technical characteristics. The inclusion of this

dataset was particularly important, given its

Data-Driven Analysis of Bicycle Lane Safety in Mexican Cities: Towards a Real-Time Route Recommendation System for Cyclists

545

focus on bicycle infrastructure, offering vital

information to assess areas with higher cyclist

risk in urban settings.

These databases were selected because of their

thoroughness and the fact that they are publicly

accessible within Mexico. They provided detailed

information on traffic accidents, with a particular

emphasis on cyclist involvement, which is critical for

evaluating the safety of cycling routes in urban

environments using both dedicated bike lanes and

conventional traffic lanes.

The analysis of these datasets allowed for the

identification of patterns that will support the

development of strategies aimed at improving urban

mobility and reducing accident rates related to

cycling. For this preliminary analysis, the data were

processed using Python to detect patterns and suggest

specific locations on the map interpreted as high-risk

and low-risk accident zones. Finally, the results were

visualized using Geographic Information System

(GIS) software, including QGIS (freeware,

https://www.qgis.org) and OpenStreetMap (freeware,

https://www.openstreetmap.org).

3 DEVELOPMENT AND FIRST

RESULTS

This section summarizes the development actions

corresponding to the first stage of the project. These

actions include the integration, processing, and

visualization of the three selected databases; the

processing and visualization of hot spots; and the risk

weighting of cycling routes.

3.1 Integration, Processing, and

Visualization of the Selected

Databases

The integration process began with an extensive pre-

processing stage, involving the following actions:

1. Format Normalization: it consisted in

standardization of geographic coordinates to a

unified system (EPSG:32613), the

homogenization of temporal formats, and

unification of accident and vehicle type

nomenclatures

2. Data Cleaning: This step was characterized by

the identification and correction of outliers, the

validation of geographic coordinates, and the

consistency checks for key fields.

Subsequently, the databases were unified through

a Python-based algorithm, implementing the

following steps: a) Identification of common fields;

b) Standardization of field names; c) Duplicate

identification based on spatial and temporal

proximity.

Duplicate records were identified based on

spatiotemporal criteria, considering records as

duplicates if they met the following conditions: a)

Spatial distance of less than 3 meters; b) Temporal

difference of less than 1 hour; c) Matching accident

type and involved vehicles.

The integration process resulted in a unified

database containing 829 filtered bicycle accident

records, with a temporal coverage from January 2015

to June 2024. The results of this initial data

processing and representation are shown in Figure 1.

Figure 1: The 829 filtered bicycle accident records.

3.2 Processing and Visualization of Hot

Spots Processing

The spatial analysis of hotspots was executed using

QGIS 3.34, implementing a multiscale method for the

generation of the heat density map. The configuration

parameters were set according to specific technical

criteria:

 Radius of influence: 20 meters (determined by

standard safety braking distance).

 Maximum value: Automatic configuration

based on data distribution.

 Chromatic gradient: Turbo spectrum.

 Rendering resolution: Medium.

The application of this analysis allowed the

identification and categorization of critical zones,

resulting in three high-risk and seven moderate-risk

areas. Figure 2 presents the spatial visualization of

these critical points resulting from the analysis.

VEHITS 2025 - 11th International Conference on Vehicle Technology and Intelligent Transport Systems

546

3.3 Risk Weighting of Cycling Routes

A Python algorithm was developed to assign risk

weights to cycling route segments, considering the

following criteria:

 Spatial proximity to the cycling route segment.

 Temporal frequency of incidents.

The initial results of this processing are presented in

Figure 3.

Figure 2: Map representation of the three high-risk areas

and seven moderate-risk areas (Note: Hotspots were

highlighted manually in this figure to facilitate their

identification within the article).

Figure 3: Map representation of the weighted risk bicycle

routes (Note: Weighted lanes were highlighted manually in

this figure to facilitate their identification within the

article).

4 CONCLUSIONS AND FUTURE

WORK

Obtained results demonstrate the effectiveness of the

proposed methodology for identifying high-risk

zones in bicycle lanes. The integration of multiple

data sources allowed for a more precise

characterization of accident patterns, while the

geospatial analysis facilitated the objective

identification of critical areas. This Data-Driven

Analysis strategy implemented this strategy offers

several advantages compared to related works,

including a) More accuracy in identifying risk zones;

b) A replicable method for database integration in

different urban contexts; and c) A first approach to

the development and implementation of specific

algorithms for cyclist accident analysis was

summarized.

Thus, we infer that these preliminary results could

serve as a reliable foundation for the development of

intelligent technological solutions that provide real-

time safe route recommendations for cyclists. This

could significantly improve urban cycling safety in

Mexican cities and, in the medium term, contribute to

secure urban eco-mobility strategies with real-world

applicability across much of the national territory.

The next stages of the project involve exploring

machine learning techniques for data processing to

translate current analysis and results to real−time

route recommendations for cyclists.

It is important to note that the code generated as

part of the project is intended to be fully accessible;

however, at this initial stage, it is not yet available as

it is still undergoing testing and will be part of a

broader solution. Eventually, much more details on

the algorithms developed to assign risk weights to

cycling route segments, and the full code will be

made publicly available to promote collaboration and

knowledge generation.

REFERENCES

Al-Ansari, H. A., & Al-Khafaji, A. S. (2024). The Role of

Safety and Security in Achieving Urban Space Vitality.

International Journal of Safety & Security Engineering,

14(2).

Avina-Bravo, E. G., Cassirame, J., Escriba, C., Acco, P.,

Fourniols, J. Y., & Soto-Romero, G. (2022). Smart

electrically assisted bicycles as health monitoring

systems: a review. Sensors, 22(2), 468.

Bahmankhah, B., & Coelho, M. C. (2017). Multi-objective

optimization for short distance trips in an urban area:

choosing between motor vehicle or cycling mobility for

a safe, smooth and less polluted route. Transportation

Research Procedia, 27, 428-435.

Batool, T., Ross, V., Blerk, J. V., Neven, A., Dendale, P.,

Janssens, D., Wets, G., & Brijs, K. (2024). Promoting

sustainable transportation: a transtheoretical

examination of active transport modes. Sustainability,

16(2), 472.

Černá, A., Černý, J., Malucelli, F., Nonato, M., Polena, L.,

& Giovannini, A. (2014). Designing optimal routes for

cycle-tourists. Transportation Research Procedia, 3,

856-865.

Chavez, C.V., Ruiz, E., Rodríguez, A.G., Pena, I.R., Larios-

Rosillo, V.M., Villanueva-Rosales, N., Mondragon,

Data-Driven Analysis of Bicycle Lane Safety in Mexican Cities: Towards a Real-Time Route Recommendation System for Cyclists

547

O.A., & Cheu, R.L. (2019). Towards Improving Safety

in Urban Mobility Using Crowdsourcing Incident Data

Collection. 2019 IEEE International Smart Cities

Conference (ISC2), 626-631.

Etminani-Ghasrodashti, R., Paydar, M., & Ardeshiri, A.

(2018). Recreational cycling in a coastal city:

Investigating lifestyle, attitudes and built environment

in cycling behavior. Sustainable cities and Society, 39,

241-251.

Ferreira, J.M., & Costa, D.G. (2024). Enhancing Cycling

Safety in Smart Cities: A Data-Driven Embedded Risk

Alert System. Smart Cities, 1992-2014.

Gulati, V. (2024). Designing of an Eco-Friendly Foldable

Bicycle using basic concepts of engineering: E-

Compaq Bike. Journal of Engineering and Technology

Management, 74, 299-304.

Heesch, K. C., & Sahlqvist, S. (2013). Key influences on

motivations for utility cycling (cycling for transport to

and from places). Health promotion journal of

Australia, 24(3), 227-233.

Khademi, N., Naeinizadeh, M., Yeganeh, S. F.,

Behmanesh, A., Ekramifard, A., Chaiyasarn, K.,

Zheng, Z., Arbabpour Bidgoli, M., Azarmi, H.

Tarvirdizadeh, B., & Hadi, A. (2024). Building a less

intimidating cycling environment for women: A

structural equation modeling analysis based on a VR-

based laboratory experiment. Transportation research

part F: traffic psychology and behaviour, 100, 431-457.

Kočková, J., Dytrt, Z., Šimeček, M., & Adamovská, E.

(2024). Sharing the Roads: Exploring Drivers' Attitudes

and Cyclists' Perceived Legitimacy in the Czech

Republic. Available at SSRN 4681768.

Komarica, J., Glavić, D., & Kaplanović, S. (2024).

Predicting and Analyzing Electric Bicycle Adoption to

Enhance Urban Mobility in Belgrade Using ANN

Models. Applied Sciences, 14(19), 8965.

Koska, T., Schneider, P., Wetzchewald, A., & Ramesohl, S.

(2021) The Path to Sustainable Mobility Systems. A

study commissioned by Huawei Technologies

Germany GmbH. Available at URL: https://web.archi

ve.org/web/20220516082643id_/https://epub.wupperi

nst.org/frontdoor/deliver/index/docId/7939/file/7939_

Mobility_Systems.pdf

Kosmidis, I., & Müller-Eie, D. (2024). The synergy of

bicycles and public transport: a systematic literature

review. Transport reviews, 44(1), 34-68.

Lagunes-Millan, M (2018) La ciclovia como movilidad

sustentable; una propuesta mediante el análisis espacial

con geotecnologías, caso de estudio Zona Urbana de

Toluca, Master’s Thesis, Universidad Autónoma del

Estado de México. Available at URL:

http://ri.uaemex.mx/bitstream/handle/20.500.11799/99

443/MILLANLAGUNASMICHAEL-MC-MAEGI-

2018.pdf?sequence=1

Liu, Y., Maurer, M. L., Carstensen, T. A., Wagner, A. M.,

Skov-Petersen, H., & Olafsson, A. S. (2024). An

integrated approach for urban green travel

environments: Planning factors, benefits and barriers as

perceived by users and planners. Journal of Transport

Geography, 117, 103849.

Mahfouz, H., Lovelace, R., & Arcaute, E. (2023). A road

segment prioritization approach for cycling

infrastructure. Journal of transport geography, 113,

103715.

Nasir, R., Meng, H., Ahmad, S. R., Waseem, L. A., Naqvi,

S. A. A., Shahid, M., ... & Tariq, A. (2024). Towards

sustainable transportation: A case study analysis of

climate-responsive strategies in a developing nation.

Case Studies in Thermal Engineering, 55, 104117.

Ogryzek, M., Adamska-Kmieć, D., & Klimach, A. (2020).

Sustainable transport: An efficient transportation

network—Case study. Sustainability, 12(19), 8274.

Peralta, M. D. (2016) Enseñanzas del Sistema EcoBici para

la implementación de Sistemas de Bicicleta Pública en

México. Friedrich-Ebert-Stiftung in Mexico. Available

online at: URL https://library.fes.de/pdf-

files/bueros/mexiko/13013.pdf

Pindarwati, A., & Wijayanto, A.W. (2019). Personalized

safety route recommender application using integrated

urban crime hotspot data. Sixth Geoinformation

Science Symposium.

Seoudi, M.S., Mesabah, I., Sabry, M., & Soubra, H. (2023).

Virtual Bike Lanes for Smart, Safe, and Green

Navigation. 2023 IEEE Conference on Power

Electronics and Renewable Energy (CPERE), 1-6.

Toski, M., Cuevas, E., Avila, K., & Perez-Cisneros, M.

(2024). Enhancing Bicycle Trajectory Planning in

Urban Environments through Complex Network

Optimization. Journal of Urban Planning and

Development, 150(3), 04024023.

Useche, S. A., Alonso, F., Boyko, A., Buyvol, P.,

Makarova, I., Parsin, G., & Faus, M. (2024). Promoting

(Safe) Young-User Cycling in Russian Cities:

Relationships among Riders’ Features, Cycling

Behaviors and Safety-Related Incidents. Sustainability,

16(8), 3193.

Visser, W. (2017). Our common future (‘The brundtland

report’) world commission on environment and

development (1987). In The top 50 sustainability books

(pp. 52-55). Routledge.

Younus, M., Nurmandi, A., Mutiarin, D., Prianto, A. L.,

Karim, A. R. M., Manaf, H. A., Bambang, I., Sohsan,

I., Mappasere, F. A., & Gang, L. (2025). Exploring the

Digital Innovation in Smart City Through Smart Public

Transportation. In J. Mukthar K. P., V. Jain, S. Bing

Tsai, C. Huei Wu (Eds.) “Digital Innovations for

Renewable Energy and Conservation”, Chapter 7, 141-

166, IGI Global.

VEHITS 2025 - 11th International Conference on Vehicle Technology and Intelligent Transport Systems

548