Impact of COVID-19 on bicycle sharing system in Oslo

Abstract

The impact of COVID-19 on Oslo bicycle sharing system was analyzed in this thesis. Weather data in the form of temperature, precipitation and wind speed, and calendar event factor (whether the day, when the ride was taken, was workday or weekend) were used as explanatory variables to assess shared bikes usage. Independent variables were chosen using LASSO shrinkage method. Response variables were daily rides and average trip duration. Predictions of sharing system usage, if no COVID happened, were made using generalized additive model. Afterwards, predictions and actual values were compared to estimate the effect of COVID-19 on Oslo system and suggestions for improvements of Oslo City Bike were given. Weather data and bicycle usage data were obtained from the open sources managed by the official providers of this information such as Norwegian Meteorological Institute and Oslo City Bike respectively. R was used to analyze this pool of data. Data were divided for train (pre pandemic period) and test (pandemic period). Based on conducted analysis, it was concluded that overall daily rides on shared bikes in Oslo decreased and average trip duration increased due to COVID-19. Number of rides also could have been affected by the appearance of massive amount of electric scooters on the streets of Oslo in 2019-2021. Oslo bicycle sharing system was able to satisfy the decreased aggregate demand for rides, but in terms of satisfying the demand for longer trips the supply side was limited to Ring 3 (road #150) of Oslo. This could be improved by rearranging placements of existing stations at a lower total cost or by expanding the bicycle sharing system’s coverage area with new bikes and new stations at a higher total cost.