Data Preparation Guide

Getting your data just right for Perceptif may not be an obvious process but is the most important part of any analysis. This guide will list out a few pointers to get the most out of your data.

📘 General Requirements

For this section we assume that there is an existing IT system in place that can export its data

For Process Mining there are two requirements that the data must already possess to form an event log

1. Is the data Structured?

   1. The data must be structured, in rows and columns.

2. Is there a Identifier, Activity and Timestamp?

   1. Otherwise called Case ID, the data may have multiple Identifiers and can retain all of them as each Identifier can be selected to give a different perspective

   2. Activities are the actions of the process and should be human-readable e.g. “Invoice Sent” instead of “1”

   3. One timestamp is required but both a start and an end timestamp columns are recommended for a more accurate analysis.

      1. Having time that includes seconds also improves the analysis.

Further Questions

1. Is your data in one file?

   1. Currently Perceptif only accepts one source of data per stream therefore external systems should be used to merge the data in the event where there are multiple files.

2. What if my system can’t export this data?

   1. By understanding the activities in a business process and then looking at the data that does exist for the timestamps and IDs corresponding to those activities. The event log can be reconstructed.