The Dutch railway network, which is very sensitive to disruptions, is amongst
the busiest ones in the European Union. In this paper, we aim to estimate the
effects of extreme weather on disruptions in the railway sector and the effect
of weather conditions on train punctuality in the Netherlands. We use a data set
from ProRail, the management of the Dutch railway infrastructures for the period
1999-2008 which includes 424,768 disruptions and the daily punctuality data from
2001 to 2010, and meteorological data for 36 weather stations provided by the
KNMI (The Royal Netherlands Meteorological Institute). We build a linear
regression model using weather measurements such as wind, lightning, temperature
and precipitation to see their impact and to predict the number of disruptions
due to weather. We also build linear regression models to find the overall
effect of weather on punctuality with and without distinguishing whether it is
due to infrastructure or due to mismanagement of the trains.