Public Bicycle Use in Paris
How can we understand and compare patterns of human movement? Are individuals moving more or less than normal at the moment? And where in Paris are people moving?
We estimate one form of active travel – use of the Vélib’ – by looking at how numbers of bicycles at docking stations change. By combining measures of docking station activity with quarters of Paris we can estimate where and when travel by hire bicycle occurs.
By comparing daily use with that expected between January 6th and March 1st 2020 (“anomaly” calculated with the signed chi score), we can begin to see the effect of the Covid-19 crisis on movement patterns – where and when we see more or less bicycle activity than expected. Expected use is calculated by day of week for each local region so we can account for expected weekday/weekend fluctuations.
Monthly patterns in 2020:
Variation by Locality
We can compare movement in local regions: through mouseover the values for an individual day, the expected value and the anomaly are shown. With a mouse click on a cell, a single locality can be selected and is highlighted in all three visualisations. Shift-click allows selection of multiple localities, and double-click selects all of Paris.
Stations are sorted vertically by the overall volume of activity with busier areas towards the top.
The vertical patterns in the charts show clearly how mobility in Paris before, during and after the lockdown changed and that weekday and weekend changes differ. We also see the effect of public holidays (especially Easter) on, what is in many cases, discretionary travel.
Team
Conceived and implemented by
- Ross Purves – Department of Geography, University of Zurich
- Ralph Straumann – EBP, Data Science Team, Zurich
- Jo Wood – giCentre, City, University of London
Data preparation by