news
Survey uses machine-learning to rank best cities in France for cycling
The new study ranked cycling conditions across France’s biggest cities
A new ranking has revealed which of France’s largest cities offer the safest and most comfortable conditions for cyclists.
The research looks at cycling conditions across the 30 biggest cities of France, including the quality of cycle routes and the conditions cyclists encounter on the road.
It was carried out by Tristan Lemoalle, an engineer and doctoral student at ENTPE, the French school specialising in sustainable development, transport and urban planning.
The study is notable because it assesses conditions at street level, instead of giving each city an overall score based on cyclists’ opinions.
Nîmes (Gard) topped the ranking, scoring 3.89 out of five for cycling conditions, ahead of Reims (Marne) on 3.77 and Paris on 3.66.
At the other end of the table, Limoges (Haute-Vienne) came last, with a score of 2.85 out of five. Toulouse (Haute-Garonne) scored 2.96, while Clermont-Ferrand (Puy-de-Dôme) recorded 3.05.
However, Mr Lemoalle said the aim was not to criticise cities, but to give them more precise information about where cycling conditions could be improved.
“The idea is not to point the finger at certain cities, but to help them better understand how users experience their infrastructure and provide precise information for creating longer-lasting and more inclusive cycling facilities.”
How were the cities assessed?
The ranking was produced using a machine-learning system that combines data on cycle lanes, traffic, road gradients and the surrounding environment to assess conditions street by street.
The system can then produce a detailed map showing cycling conditions and highlighting sections where improvements may be needed.
The research is intended to be used by local authorities as well as for producing the national ranking.
The programme can provide a city with an overall score as well as a detailed map showing the strengths and weaknesses of its cycling network.