How to provide intelligent lighting in Paris streets by analysing mobile and urban travel data?
THE CONTEXT
In Paris, as in many other cities, street lights remain lit with the same level of brightness day and night, although they could be dimmed during less frequented times or brightened for specific events (around stadiums on match nights) in order to reduce carbon footprint and light pollution. This intelligent street lighting system is particularly important for soft transport needs (walking, cycling, etc.).
3% to10%
potential savings on the city’s annual
electricity bill for street lighting
(average €1 million/year)
THE CHALLENGE
Pilot projects with sensors and meters on individual lights have proven to be costly, and with unreliable results. The objective of this challenge was to use geostatistics to adapt street light to actual needs. Partners provided the following data sets: street lighting facilities data as well as anonymised geo-located data from mobile devices.
THE STARTUPS
Two startups teamed up to tackle this challenge, Quantmetry a data science consulting company and Dataiku providing a collaborative data science platform to prototype and deploy solutions.
THE PARTNERS
THE SOLUTION
The two startups teamed up to model urban travels at night, crossing anonymised and aggregated data from SFR’s mobile network, and urban streets and travel data coming from the open database of the city of Paris. They demonstrated the possibility to adapt street lighting according to urban travels. A web app was developed, to enable the visualisation of off-peak times when the lights can be dimmed, as well as cost and energy savings related to the dimming enabling the city of Paris to reduce by 3-10% its yearly electricity bill. This solution tested in the 13th district is scalable to the 199 000 street lights in Paris and to any street light system in the world using mobile data, without installing any specific sensors or equipment.