Process mining techniques such as process discovery and conformance checking provide insights into actual processes by analyzing event data that are widely available in information systems. These data are very valuable, but often contain sensitive information, and process analysts need to balance con dentiality and utility. Privacy issues in process mining are recently receiving more attention from researchers which should be complemented by a tool to integrate the solutions and make them available in the real world. In this paper, we introduce a Pythonbased infrastructure implementing state-of-the-art privacy preservation techniques in process mining. The infrastructure provides a hierarchy of usages from single techniques to the collection of techniques, integrated as web-based tools. Our infrastructure manages both standard and nonstandard event data resulting from privacy preservation techniques. It also stores explicit privacy metadata to track the modi cations applied to protect sensitive data.
Process mining provides fact-based insights into actual business processes using
event data, which are often stored in the form of event logs. The three basic
types of process mining are process discovery, conformance checking, and
process enhancement [
Privacy issues in process mining are highlighted when the individuals' data
are included in the event logs. According to the regulations such as the European
General Data Protection Regulation (GDPR) [
Figure 1 shows the general approach of privacy in process mining including tDwaotamPaiunblaischtiivnigtie(sP:PPDriPv)acayn-dPrPesreivravicnyg- Eve(nEtL)Log PrDivaatacyP-Purbelissehrivnigng AbstErvaecntitoLno(gELA) Preserving Process Mining (PPPM). EL PPDP ELA PPDP aims to hide the identity and Privacy-Aware tecnPhevoPseensPn-smtsMetnaidnnsaaiidinttmiagavrseadtltgdoodoaaretptiaxatrthoeormtenfesdscrutetltoctriotanwhrdgdoeiritfrkioroowwnpmnairtelihsvrposatmrchoiynee-. iirsscgePnnoM PM EventELLo'gPM(EiirsscgePnnoML') PPPM ii-rrrscvvgyaeePPn iirsscgePnnoM gPoPrDithPmtsecahrneiqtuigesh.tlNyocteoutphlaetdPwPiPthMtahle- ProRceessuslMt(iRn)ing R≈R' ProRceessusltM(Rin')ing ProRceessusltM(Rin')ing corresponding PPDP techniques.
In this paper, we introduce a tool Fig. 1: The general approach of privacy in
which mainly focuses on PPDP and process mining.
o ers state-of-the-art privacy
preservation techniques including the connector
method for securely discovering processes [
The remainder of the paper is organized as follows. In Section 2, we demonstrate the functionality and characteristics of the tool. Section 3 outlines the maturity and availability of the tool, and Section 4 concludes the paper. 2
In this section, we demonstrate the main functionalities and characteristics of
our stand-alone web-based tool, PPDP-PM, which is written in Python using
Django framework1. Our tool has four main modules: event data management,
privacy-aware role mining, connector method, and TLKC-privacy. The event data
management module has two tabs to upload and manage the event data that
could be standard XES event logs2 or non-standard event data, called Event
Log Abstraction (ELA) [
Add an output to theevent datamanagementmodule.
Link to the GitHub project.
Outputs are temporarily
stored here.
for the privacy preservation techniques. The privacy-aware role mining module
(Figure 2) implements the decomposition method supporting three di erent
techniques: xed-value, selective, and frequency-based [
The connector method implements an encryption-based method for
discovering directly follows graphs [
Each privacy preservation technique in the tool is implemented as a Django application that enables the simultaneous running of di erent techniques on an event log. This architecture makes the whole project easy to maintain, and new techniques can simply be integrated as independent applications. The outputs for the privacy preservation techniques are provided independently for each technique and can be downloaded or stored in the event data repository. PPDP-PM is designed in a way that provides a cycle of privacy preservation techniques, i.e., the privacy-aware event data, added to the event data repository, can be set as the input for the techniques again as long as they are in the form of standard XES event logs. To keep the process analysts aware of the modi cations applied
Process miningalgorithms thatcanbe appliedtothe results fromthe privacy preservationtechniques.
Privacypreservation
techniques integrated
inPMTK.
to the privacy-aware event logs, the privacy metadata [
As mentioned, PPDP-PM is a web-based application written in Python. The source code, a screencast, and other information are available in a GitHub repository: https://github.com/m4jidRafiei/PPDP-PM. The privacy preservation techniques, explained in Section 2, and the integration into PMTK are also available as separate GitHub repositories.3 To facilitate the usage and integration of the privacy preservation techniques, they are also published as standard Python packages (https://pypi.org/): pp-role-mining , p-connector-dfg , ptlkc-privacy , and p-privacy-metadata. Our infrastructure provides a hierarchy of usages such that users can use each technique independently, they can use PPDP-PM which integrates a set of privacy preservation techniques as a standalone web-based application, and they can also use the provided techniques in a process mining tool where the privacy preservation techniques are integrated. The scalability of the tool varies w.r.t. the privacy preservation technique and the size of the input event log. Based on our experiments, our tool can handle real-world event logs, e.g., the BPI challenge datasets4. However, it can still be improved for industry-scale usage. PPDP-PM and its integration in PMTK are also provided as Docker containers which can simply be hosted by the users: https://hub.docker.com/u/m4jid. 4
Event data often include highly sensitive information that needs to be considered by process analysts w.r.t. the regulations. In this paper, we introduced a Python-based infrastructure for dealing with privacy issues in process mining. A web-based application was introduced implementing privacy-preserving data
publishing techniques in process mining. We also showed the privacy integration in PMTK as an open-source web-based process mining tool. The infrastructure was designed in such a way that other privacy preservation techniques can be integrated. We plan to cover di erent perspectives of privacy and con dentiality issues in process mining, and novel techniques are supposed to be integrated into the introduced framework. We also invite other researchers to integrate their solutions as independent applications in the provided framework.