BPI Challenge 2014 — Get Started Now!

Process Mining Challenge 2014 - Start now!

The BPI Challenge is an annual process mining competition, which takes place for the fourth time this year. The goal of the challenge is to give both researchers and practitioners the opportunity to do process mining analyses on real-life data (read our interview with Boudewijn, where he tells the story of how the BPI Challenge came to life).

In this competition, anonymized but real data is provided and can be analyzed by anyone using any tools. Submissions can be handed in until July 12, 2014 and the winner will be awarded a prestigious prize!

As always, we make our process mining software Disco available for anyone for the purpose of this challenge. Read on to see what this year's challenge is about and how you can get started.

The Process

This year's data is provided by the Rabobank Group ICT.

Similar to other ICT companies, Rabobank Group ICT has to implement an increasing number of software releases, while the time to market is decreasing. Rabobank Group ICT has implemented the ITIL-processes and uses the Change process for implementing these so called planned changes.

The data is provided for the following service desk processes: Interaction Management, Incident Management, Change Management (see below).

Process BPI Challenge 2014 (click to enlarge)

As you can see in the illustration, a problem reported to the service desk (for example, a slow internet connection) may evolve from an Interaction (an agent from the service desks troubleshoots the reported issue) to an Incident (the issue cannot be solved on the phone but someone has to look into it) up to a Change request (repeated problems of the same kind lead to a structural change that should prevent issues in the future).

The following detailed information is provided about each stage.

Interaction Management

In order to manage calls or mails from customers (Rabobank colleagues) to the Service Desk concerning disruptions of ICT-services, a Service Desk Agent (SDA) logs calls/mails in an Interaction-record and relates it to an Affected Configuration Item (CI). The SDA can either resolve the issue for the customer directly (First Call Resolution) or create an Incident-record to assign the issue to an Assignment Group with more technical knowledge to resolve the service disruption. If similar calls/mails are received by the Service Desk, a SDA can decide to relate multiple Interaction-records to one Incident-record. Further logging of Activities to resolve the service disruption will be done in the Incident-record.

Incident Management

Based on an estimated Impact and Urgency, done by the SDA, an Incident-record is prioritized and gets a deadline to resolve the service disruption. A Team leader within the Assignment Group assigns the records to an Operator. The Operator resolves the issue for the customer, or reassigns the record to a colleague if other or more knowledge is needed. After solving the issue for the customer, the Operator relates the Incident-record to the Configuration Item (CausedBy CI) that caused the service disruption. After closing the Incident-record, the customer receives an email to inform that the issue is resolved.

Change Management

If particular service disruptions reoccur more often than usual, a Problem investigation is started, which will lead to an analysis and improvement plan to prevent the service disruption to happen again. The improvement plan leads to a Request for Change (RfC) on the CausedBy CI. All CIs are related to a Service Component, Risk Impact Analysis is done by an Implementation Manager assigned to changes related to the specific Service Component.

The Data Set

As in any process mining analysis, the data needs to be linked to a Case ID, Activity, and Timestamp.

The data that you will analyze in this BPI Challenge stems from the IT Service Management (ITSM) software that is used in the service desk to handle the processes described above. Activities and timestamps are recorded within the ITSM system for the processed interactions, incidents, and changes.

An additional difficulty this year is that the data is provided in four pieces for the different processes. For the process mining analysis, the data needs to be combined and this can be done in different ways. It is part of the analysis to understand and prepare the data according to the questions and goals of the analysis.

If you click on the picture below, you can see the data fields that are contained in the four files:

Data model of provided data (click to enlarge)

We have imported these four files and already created two additional views for you in a Disco project file that you can simply open with the freely available demo version of Disco.

You can download both the Disco project file and the raw data and data model explanation here:

Download the raw data files and the data model in a Zip file Download the raw data files in a Zip file (CSV files and explanation reference about the data model)

BPIC-2014.zip

Download the Disco project file that can be opened with the demo version of Disco Download the Disco project file that can be opened with the freely available demo version of Disco

BPI-2014.dsc

We think that many of you will want to create additional views by combining or importing the data in different ways. The two views that we created (integrated incidents and a more detailed view on the change process) are just an example and a starting point.

If you have created another combination of the data that you want to analyze in Disco as well, you can send the file to anne@fluxicon.com and we will add your view to the project file.1 You will be mentioned as the creator of the new data view and updates to the project file will be made available here on this blog post for the community.

New This Year: Student Challenge

View Process Mining Research Institutes in a larger map

Students all over the world are starting to learn about process mining. However, learning about it in theory and applying process mining in practice are quite a different story.

To give students the possibility to develop their process mining skills, this year's BPI Challenge, for the first time, includes a separate student competition. Student groups are invited to participate in the challenge, and their submissions will be evaluated separately from the regular submissions.

Because students normally do not have any experience with process analysis and improvement work at companies, we decided to pair them with a mentor who is a practitioner and can guide them and be available for questions. This way, the student teams will learn more and deliver better analyses. More than a dozen practitioners in Europe, Scandinavia, the US, and South America have already volunteered to mentor a student team. To be matched with a mentor, students can email anne@fluxicon.com by 31 May 2014.

_Update: Please apply to the mentorship program through the form at the BPIC 2014 Student Challenge website._

Do you want to be a mentor as well? This in no way hinders your own participation in this year's BPI challenge (the student challenge is completely separate). Let us know and we will try to match you with a student team in your geographical region.

Now, there is one more bonus attached to the student competition: Two extra prizes are available for the winners.

Questions About the Process

One of the challenges of a process mining project is that you need a starting point to understand the process context and what business questions and goals are relevant for the analysis. Otherwise it is really easy to get lost in the data.

This is why the data in the BPI Challenge is not just dropped over the fence, but the data providers are encouraged to provide questions that can be the starting point for the people who are participating in the challenge.

Here is what the Rabobank ideally would like to know about the data:

Rabobank Group ICT is looking for fact-based insight into sub questions, concerning the impact of changes in the past, to predict the workload at the Service Desk and/or IT Operations after future changes.

The challenge is to design a (draft) predictive model, which can be used to implement in a BI environment. The purpose of this predictive model will be to support Business Change Management in implementing software releases with less impact on the Service Desk and/or IT Operations.

We have prepared several case-files with anonymous information from Rabobank Netherlands Group ICT for this challenge. The files contain record details from an ITIL Service Management tool called HP Service Manager. We provide you with extracts in CSV with the Interaction-, Incident- or Change-number as case ID. Next to these case-files, we provide you with an Activity-log, related to the Incident-cases. There is also a document detailing the data in the CSV file and providing background to the Service Management tool.

Sub questions

  1. Identification of Impact-patterns: We expect there to be a correlation between the implementation of a change and the workload in the Service Desk (SD) and/or IT Operations (ITO), i.e. increased/decreased volume of Closed Interactions and/or increased/decreased volume of Closed Incidents. Rabobank Group ICT is interested in identifying any patterns that may be visible in the log for various service components to which a configuration item is related, in order to predict the workload at the SD and/or ITO after future changes.

  2. Parameters for every Impact-pattern: In order to be able to use the results of prior changes to predict the workload for the Service Desk directly after the implementation of future changes, we are interested in the following parameters for every impact-pattern investigated in sub question 1:

* What is the average period to return to a steady state?
* What is the average increase/decrease of Closed Interactions once a new steady state is reached?
  1. Change in Average Steps to Resolution: Since project managers are expected to deliver the same or better service levels after each change implementation, Rabobank Group ICT is looking for confirmation that this challenge is indeed being met for all or many Service Components.

  2. Creativity challenge: Finally, we challenge the creative minds, to surprise Rabobank Group ICT with new insights on the provided data to help change implementation teams to continuously improve their Standard Operation Procedures.

What can you do if these questions are not interesting or not feasible for you? After all, you may need quite some data mining skills to fully address the questions above.

Don't worry. Like explained above, the questions from the process owner are intended to provide you with a starting point. The BPI Challenge gives you the chance to practice your process mining skills on real data and there are many ways to do this. Think of a process question that would be relevant for you, for your clients, or - if you are a researcher - what insights would your fantastic new algorithm add in this situation?

What is important is that you clearly state the questions and the reasoning behind your analysis. Motivate why the question is relevant and describe how you approach the analysis in sufficient detail, so that others can understand what you did and why.

How to Submit

You can submit your challenge contribution through the EasyChair system at https://www.easychair.org/conferences/?conf=bpic2014.

A submission should contain a pdf report of at most 30 pages, including figures, using the LNCS/LNBIP format specified by Springer (available both as a Word and as LaTeX template). Appendices may be included, but should only support the main text.

Submission deadline: July 12, 2014, 23:59 CET

Announcement of winners: at the 10th Workshop on Business Process Intelligence (BPI 14), Haifa, Israel, 8th September 2014

Questions?

Join us for a webinar, where we have invited both the challenge organizers and a process expert from the Rabobank. This is your chance to answer all your questions about the challenge and about the data set.

The tentative date for the webinar is 15 May at 17:00 CET. Sign up now to make sure you don't miss it!


  1. If you have a license for Disco and are not working with the demo version, of course you can simply import the data yourself.

Anne Rozinat

Anne Rozinat

Market, customers, and everything else

Anne knows how to mine a process like no other. She has conducted a large number of process mining projects with companies such as Philips Healthcare, Océ, ASML, Philips Consumer Lifestyle, and many others.