Analysis of freight traffic flows in R Shiny

Challenge

European freight transport is conducted through various modes, such as rail, truck, and inland shipping. To identify growth potentials and possible competitive situations, different data are fed into a forecasting model. This information, from both internal and external sources, needs to be harmonized and structured uniformly for the forecasting model. The subsequent evaluation of the results and communication to decision-makers also requires a standardized analysis tool.

Approach

In the initial steps, the previously used scripts for data preparation were fundamentally revised, the data basis was expanded, and generalized logics were implemented and documented. Performance issues and unnecessary interdependencies between different preparation steps were eliminated. Subsequently, all scripts were functionalized and published in an internal R package. This allows for clear versioning and communication of changes, as well as easy use and execution of the scripts by many less familiar users. The data processable by the forecasting model is stored in a memory-efficient database. An R Shiny application was developed for analyzing forecasts, enabling quick and consistent evaluation of the impacts of new data or other parameter constellations.

Result

In addition to optimized data quality, the improved understanding of individual steps significantly contributes to better forecasting. The graphical and tabular analyses implemented in the application make the evaluation much more structured and faster. Presentations-ready visuals can be generated automatically.