Recipients: University of Seville. 2018.
From open data published by the Sevici’s management company and compiled by the Universities of Huelva and Seville for a year, the temporal, spatial and spatio-temporal patterns of the use of Sevici’s bicycles are identified. The results confirm their effective use as a means of transport (alternative) in daily activities and also, intuit the preferred functional role in the urban structure of the different areas in which the stations are located.
Predictive models of the use of bicycles have also been built, for short-range time horizons (15min, 30min, 1h, 4h, 8h and 24h). In total, 1813 models have been estimated, seven for each of the stations, all, regression models with elasticnet regularization. The results of their testing have shown a very good predictive capacity, with RMSE (root mean square error) of 17.6 on average (percentage of available bikes).