Mapping datasets

Mapping Datasets in ProcessMind

Mapping datasets is a crucial step in turning raw data into insights within ProcessMind. One of the key strengths of ProcessMind is its flexibility: data can always be added, removed, enabled, or disabled as you go. There’s no fixed order or forced way to integrate data into your process.

You have two primary approaches to working with data:

1. Adding Data After Defining a Process: Start with a process framework and incorporate data to identify missing information and refine your analysis.
2. Starting with Data: Use your event data to create an initial process map and build from there.

Which Approach Should I Pick?

Situation	Approach
You already have a defined process and want to enrich it with data	Start with a process, then add data
You only have raw event data and no model yet	Start with data and mine the process
You want to evaluate ProcessMind first	Generate a sample process with simulated data

---

This documentation walks through starting with an empty canvas and gradually building your process and analysis step by step. If you prefer to start with an existing process, you can import an existing BPMN model, and map data to tasks and events of the imported model directly.

Step 1: Starting with an Empty Canvas

Begin by creating a new process or opening an existing one. The canvas serves as the foundation for your model, where you will map and organize datasets. If you did not already upload your data in the data section, you can also upload data directly from the process view. This can be done from the right panel: open the dataset selection area, as shown in the image below.

Select Your Dataset

Once your dataset is uploaded and processed, the system will notify you that it is ready to use. You can then select it from the Dataset List, as shown in the image above. The most recently uploaded dataset will always appear at the top.

When you hover over a dataset in the list, a tooltip will provide additional context, such as:

• Name of the dataset
• Number of rows identified
• Uploaded file size and name
• Uploaded date and time

This helps ensure you’re selecting the correct dataset for your process.

After selecting your dataset, the system will perform some basic pre-processing. This is indicated by a loading icon displayed next to the dataset name.

For the context of your process, you can rename the dataset (if desired) specifically for its use in this process, making it easier to identify later. A switch gives you the option to show or hide the dataset in the model.

Dataset Options

The Dataset Settings Menu offers several options to manage and customize your dataset effectively. Below is a breakdown of the available options:

• Edit dataset:
  Quickly access the Edit dataset option to make changes to the dataset itself. This allows you to modify or refine the dataset directly.

• Remove data from model:
  If you no longer need the dataset in your process, use the Remove data from model option. This removes all references to the dataset and clears it from your process canvas.
  Note: This action does not delete the dataset itself; it remains available in the Dataset List.

• Dataset Colors:
  Change the dataset color for better visual distinction. The selected color will also apply to activities that are derived from this dataset, making them easier to identify on the canvas.

• Auto map this dataset:
  This option attempts to automatically map the dataset activities to existing activities already represented in your process model. This saves time and helps maintain consistency.

• Remove all mappings of this dataset:
  Use this option to clear all activity mappings between the dataset and your process model. This is useful if you need to start over or make significant changes.

• Auto Layout Model:
  The Auto Layout Model option automatically arranges activities and their relationships on the canvas for improved representation and clarity.

• Merge datasets with same color:
  Merge datasets that share a color into a single model, unified under that shared color.

In addition, the dataset list row includes a Show data in model switch that lets you hide or show the activities found in the dataset that are not directly mapped in the model. Toggle this to manage the visibility of unmapped activities.

By using these options, you can manage how datasets are integrated, represented, and compared within your process models. For how the primary and comparison roles work — switching, promoting, and removing — see Primary and Comparison.

Mapping Data to the Process

Once the dataset is fully loaded, the system will automatically display the process mining result on the canvas. This initial process map is a free-floating model without any fixed attachments. To make it a part of your process design and editable, you need to map it to an existing model or activity, or convert it into a fixed model.

Fixing the Model to the Canvas

There are two main ways to fix activities from the floating model to the canvas:

1. Individually Select Activities:
   Select specific activities to map individually.

2. Bulk Selection:
   Use the selection tool or keyboard shortcuts to select multiple activities at once:

   • Shift + Mouse Drag: Draw a selection box around the activities.
   • Select All: Press Ctrl + A (Windows) or Command + A (macOS) to select all activities.

After making your selection, a new context menu will appear next to the selected activities. From here, you can choose Add to Model. This option fixes the selected activities to the canvas, allowing you to:

• Add context to the activities.
• Relate them to other attributes.
• Map or unmap additional attributes as needed.

By fixing your dataset’s process map to the canvas, you can begin enriching your process with detailed attributes, relationships, and context, turning raw data into insights.

Unmapped Activities

Unmapped activities are activities present in the dataset but not yet mapped to any attributes in the model. These represent potential gaps or elements that need further integration into your process design.

When you toggle the Unmapped Activities option on or off, the unmapped data will be visually represented with dotted lines. You can tell unmapped activities apart from mapped ones at a glance.

In the visual example, you can observe the before and after states of toggling the unmapped activities option:

• Before: Unmapped activities are hidden and not visible on the canvas.
• After: Unmapped activities appear with dotted lines, highlighting areas for potential mapping.

Using this feature, you can efficiently manage and address unmapped data, ensuring your process model is as complete and accurate as possible.

Build and Refine as You Go

Whether you start with a defined process or with raw data, you can build and refine your model as you go. The result is a complete model you can analyze and improve.