Identification of causes of unpunctuality in autumn

Challenge

• The autumn months have lower long-distance punctuality than other months, so the company's punctuality targets are missed
• The causes of this unpunctuality are unclear, so that measures to control it are insufficient
• The analysis is intended to identify possible unknown drivers in 2019 and subsequent years at an early stage

Solution

• Potential factors influencing long-distance punctuality are compiled in a driver tree
• The necessary data to quantify these factors are collected and linked in a large data set
• A linear model with regularisation identifies the most important punctuality drivers, quantifies their punctuality effect and their effect in autumn

Result

• Overview of relevant punctuality drivers available in the form of a driver tree
• This generic approach works with statistical models and is thus applicable to different timetable years and train types with little effort