Process mining is a technique that analyzes the data generated by business processes to discover, monitor, and improve them. It can help you understand how your processes work, where the bottlenecks are, how to optimize them, and how to ensure compliance and quality.

But how does process mining work? What kind of data does it need? And how can you create such data?
The answer is: you need an event log.

An event log is a collection of records that describe what happened in a business process. Each record, or event, contains information about:

• A case: this is an instance of the process, such as an order, a claim, a patient, etc.
• An activity: this is a step or task in the process, such as placing an order, approving a claim, diagnosing a patient, etc.
• A timestamp: this is the date and time when the event occurred, such as 2023-08-12 00:01:27.
• Additional attributes: these are optional data that provide more context or details about the event, such as the performer, the cost, the outcome, the location, etc.

An event log can be seen as a table where each row is an event and each column is an attribute. For example, here is a simplified event log of an order-to-cash process:

Case ID	Activity	Timestamp	Performer	Amount
1	Place order	2023-08-01 10:15:32	Customer A	100
1	Check availability	2023-08-01 10:16:45	Warehouse B	–
1	Confirm order	2023-08-01 10:17:12	Customer A	–
1	Prepare shipment	2023-08-01 10:18:34	Warehouse B	–
1	Send invoice	2023-08-01 10:19:56	Accounting C	–
1	Ship order	2023-08-01 10:20:23	Warehouse B	–
1	Receive payment	2023-08-02 11:21:45	Accounting C	–
2	Place order	2023-08-01 10:22:12	Customer D	200
…

From this event log, we can see that there are two cases (orders) in the process, each with a different sequence of activities and timestamps. We can also see who performed each activity and how much each order was worth.
By analyzing this event log with process mining tools, we can discover the actual flow of the process, measure its performance indicators (such as cycle time, throughput, waiting time, etc.), identify deviations or anomalies from the expected or desired behavior, and suggest improvements or optimizations.

But how can you create an event log for your own process?
The good news is that you probably already have some data sources that can provide event data. These can include:

• Database systems (e.g., patient data in a hospital)
• CSV files or spreadsheets
• Transaction logs (e.g., a trading system)
• Business suite/ERP systems (SAP, Oracle, Dynamics, etc.)
• Message logs (e.g., from IBM middleware)
• Open APIs providing data from websites or social media

The challenge is to extract the relevant data from these sources and transform them into an event log format that can be used by process mining tools. This may require some data cleaning, filtering, merging, or enriching steps to ensure the quality and completeness of the event log.

One of the most widely used formats for event logs is XES (eXtensible Event Stream), which is an XML-based standard adopted by the IEEE Task Force on Process Mining. XES allows you to define your own attributes and extensions for your event log, making it flexible and adaptable to different domains and scenarios. Many process mining tools support XES as an input or output format.

To convert your data sources into XES format, you can use various tools or methods depending on your needs and preferences. Some examples are:

• ProM: this is an open-source platform for process mining that offers various plugins for importing and exporting XES files from different sources.
• Disco: this is a commercial tool for process mining that allows you to import CSV files or connect to databases and automatically convert them into XES files.
• Python: this is a popular programming language that has several libraries for working with XES files, such as pm4py or xes.
• R: this is another popular programming language that has some packages for working with XES files, such as bupaR or xesreadR.

Once you have your event log in XES format, you can use any process mining tool that supports it to analyze your process and gain valuable insights.

In summary, a process mining event log is a collection of records that describe what happened in a business process. It contains information about the cases, activities, timestamps, and additional attributes of the events. You can create an event log from various data sources by extracting, transforming, and converting them into XES format. Then, you can use process mining tools to discover, monitor, and improve your process.

I hope this blog post helped you understand the concept of a process mining event log and how to create one. If you have any questions or comments, please feel free to leave them below. Thank you for reading! 😊