Joint Master Thesis RWTH PADS/PostNL Process Mining for Postal Processes

Kontakt

Dr. ir. Sebastiaan J. van Zelst

Name

Sebastiaan J. van Zelst

Wissenschaftlicher Mitarbeiter - Fraunhofer FIT

Telefon

work
+49 241 80 21926

E-Mail

E-Mail
 

Description

Postal services comprise of various, complex, digitalized processes being executed in order to transport and deliver various goods. On a daily basis, hundred thousands of packages are being picked up and delivered at various customers throughout the Netherlands. Clearly, efficient delivery of goods, and thus, efficient execution of the underlying processes is of utmost importance. Not only for postal service providers, but also for the customer, e.g. fast delivery of newly bought products and the environment, e.g. avoiding unnecessary rerouting of packages.

In order to ensure a smooth execution of the aforementioned processes, a proper synchronization and orchestration of the different processes at play is of major importance. Due to the inherent intertwining of the different underlying processes, a small bottleneck in one of these processes typically causes tremendous delays in other processes, hampering the overall efficiency of the postal service process. Hence, a clear understanding of the different processes active within postal services, as well as adequate counter measures when bottlenecks and other deficiencies are expected to occur, are key to streamline the overall postal service performance.

The different information systems, used by Post NL, allow us to track, often in great detail, the execution of the different processes at play. Consider for example, that each packet is scanned roughly 15 times on its journey from sender to receiver. As such, these information systems allow us to obtain valuable traces of event data. Recent developments in the research field of process mining, which represents a large body of data driven analysis techniques on the basis of such event data, allow us to gain detailed insights in the execution of these processes. The advantage of deriving insights in processes based on operational data relates to the fact that we observe what actually happened, i.e. as captured by the data.

In this M.Sc. thesis project, which is a joint project of PostNL, market leader of postal services in the Netherlands, and the RWTH PADS chair, the student is asked to investigate the application of process mining, in the context of the postal service’s process data. Initially, a complex data set, e.g. packet scan data, needs to be analysed. Based on the analysis, the student, in cooperation with the supervision team (both from PostNL and RWTH), will identify and investigate relevant postal-service-specific problems. Finally, the student proposes a generic solution to the aforementioned identified problems.

The M.Sc. project will be executed at the PADS chair, and at the PostNL headquarters in den Haag, the Netherlands. Within PostNL, the student will execute his/her work in the “Analytics & Decision Support” (ADS) group. ADS is a group of 15 experienced data scientists and four data engineers that supports business functions by generating actionable insights from applying advanced statistical methods (e.g., regression models, clustering, decision trees, random forest, simulation, optimization) to Big Data. The ADS team guides these data projects from the beginning (defining the initial business question) towards the end (implementation of the solution). With the vast amount of data within PostNL, there are plenty of opportunities for improving and optimizing processes and customer experiences. In the past ADS successfully completed many analytics projects, such as predicting the chance to successfully deliver a package, predicting which retail location will be too full, customer segmentation and optimizing delivery routes. The student will be involved in all steps that data science projects in PostNL go through. That is, first a scoping session is hold with the business partner; second, an analysis plan is made; third, the data are collected and prepared; fourth, the data are analyzed; finally, a report is made and presented. During the entire project, the student will work together with the ADS data scientists and there will be regular sessions with the business partner to discuss the progress.

This is a unique opportunity to work on real data, get direct exposure to industry, and, to get international working experience, in a multi-national team. Moreover, it is very likely that the tools and techniques developed in the context of this M.Sc. thesis work will be adopted in practice at PostNL! Note that, RWTH University does not provide any financial support for M.Sc. students. It is your responsibility, e.g. in consultation with the company or by means of obtaining a scholarship, to cover the envisioned expenses.

Prerequisites

Knowledge of basic computer science concepts, good programming skills (Java/Python) and an interest in theoretical and practical aspects of process mining are recommended.

Pointers

Academic Supervisor

Prof.dr.ir. Wil van der Aalst

Daily Supervision

Sebastiaan van Zelst

Company Supervision

Helen Steingröver

More Information/Application

Send an e-mail to and . Make sure to include a C.V., detailed information about your background and scores for completed courses.