Application of self-organizing map (SOM) analysis for estimating bicycle sharing: a new perspective

Meteorology may be key to forecasting whether people will use it or not (depending on city studied), but so to date, the predictions made have generated some controversy because they have not been analyzed using non-linear analysis. The objective of this study is to analyse the relationship between the time spent using the València bike sharing service (SBC) as a means of active transport and the weather. A self-organising map analysis (SOM) was performed to generate profiles (clusters) of days on SBC use and meteorological factors and a non-parametric analysis was performed to compare the different profiles generated. The results showed 8 profiles of days, which obtained multiple significant differences. These results show that, although there are variables with greater weight than others for estimating the use of the SBC, their relationship is not always linear and a combination of them is needed for greater rigor in the predictions. In this study has been observed that, in order to predict a high use of the SBC, days should be warm if humidity is low to moderate, although temperature is limited if humidity is high, with virtually no precipitation and low average wind speed. On the other hand, to estimate low SBC use, days should be characterized by high relative humidity, precipitation and wind speed. On these days, if the humidity is not high and there is no precipitation, low temperatures would be taken into account. In conclusion, the use of non-linear analyses such as SOM proves to be an effective tool for estimating the use of SBC in relation to meteorology. Keywords: Active transport, bike-sharing, SOM, meteorology, non-lineal analyses.