Introduction to Process Mining

Process mining is an analytical technique that helps to discover, monitor and improve operational business processes. Using process mining, you can see how the process actually goes, instead of assuming how the process goes. Process mining helps you to discover, explore, audit, analyze and improve your processes by answering both compliance-related and performance-related questions. The ProcessMining extension for Tableau can be offered as a on-premise (we call it Enterprise) or as SaaS (we call it Share) deployment.

To get the best value from the Process Mining extension preparing your data correctly is not difficult, but it is important. To get started with the ProcessMining extension you just connect your Tableau Desktop to a datasource containing your process event logs. These log files must contain:

• A Case identifier that identifies each unique case flowing through your process.
• An activity identifier (or “Process step”) that will be shown as the nodes in your process map
• A start Date/Timestamp
• An end Date/Timestamp (not required but really helpful)

You don’t need any measures or calculations, the most relevant process KPI’s are automatically calculated within the extension. With these 4 dimensions create a separate ‘datasheet’. It is best practice to not connect your extension to an existing visualization, use a dedicated ‘datasheet’ instead. Using a dedicated ‘datasheet’ for your ProcessMining extension will result in better performing dashboards and increased interactivity.

Creating the ‘datasheet’ is a straight forward procedure. Just drag the dimensions to the rows. Make sure your “Start” and “End” Date/Timestamps are of datatype date/time, are set to exact date and are presented as discrete values. It is recommended to use the order represented in the screenshot below.

Make sure you do a nested sort of your Case ID (and the next dimensions) on the minimum of the “Start” Date/Timestamp. This will make sure that your activities are sorted in a chronological way. The content of the ‘datasheet’ should look something like the screenshot below:

This chapter will cover all steps to create a dashboard with the ProcessMining extension for Tableau. In the previous step we explained how to prepare your data. So once your datasource is set-up and your ‘datasheet’ has been created, we start creating the dashboard.

First thing to do is create a new Dashboard. Tableau Extensions are Dashboard objects and cannot be deployed on just sheets. Therefore we need to add some data. In the previous steps we already showed you how to create a ‘datasheet’ so on the new dashboard just add your ‘datasheet’. This is a sheet that we only want to use to feed the extension so don’t mind the layout of the ‘datasheet’, we are going to hide it. For tips and advice check out our BLOG.

The datasheet is not an attractive sheet to be present on your dashboard. That’s why we are going to hide it. Remember, the datasheet must be present on the dashboard, but that doesn’t mean we can’t hide it. To hide the ‘datasheet’ just select the container it is in and open the “Lay-out” tab on the left pane in Tableau Desktop (see screenshot on the right).

Now enable the ‘Floating’ option and resize the container to 1×1 pixels (wxh) and change the position to the 1×1 x,y position. This will resize the floating container to one single pixels in the upper left corner of your dashboard (virtually invisible).

Next step is to add the ProcessMining extension to your dashboard. To complete this step you need the trex file that matches your subscription. This can either be downloaded from

1. The Tableau Extensions Gallery
2. Your personal trial download page when you are on a free 14 day trial
3. Your appsfortableau.com account when you are on a Share subscription
4. Your system administrator when you are on an Enterprise subscription

In your dashboard now drag the Extensions object to your dashboard and release it where you want to put it in your dashboard. A Tableau Modal window will pop-up that will ask you to select the extension or trex file you want to use.

In Tableau 2021.1 and higher you will see an interface where it is made easy to use a trex file from the Tableau Extensions Gallery. You can choose to select a trex file from your filesystem (marked yellow in the screenshot)

In this case we already prepared a trex file so we will use the yellow marked option and select the ProcessMining trex file. You will be asked to allow the extension in your dashboard. This message cannot be suppressed in Tableau Desktop but can be suppressed on Tableau Server / Tableau Online. Once the extension is properly loaded in your dashboard you will see the start screen that holds a getting started YouTube movie that explains the steps you need to take from here.

First we will connect the extension to our already prepared ‘datasheet’.

When adding the ProcessMining extension to a dashboard, it will open the configuration wizard by default. This wizard will take you through the initial steps of configuration:

1. Configuring the datasheet and (optionally) limiting data for testing purposes
2. Assigning the “Activity”, “Case Identifier”, “Start time” and optionally the “End time” and “Custom measure” roles to the relevant data from the data source
3. Initial configuration of the color, shapes and layout of the Process Mining visualisation

Alternatively, you can follow the steps below to configure the extensions manually, and in more detail.

This chapter will explain the steps that need to be taken to get your ProcessMining extension up and running in your dashboard. First step is to open the

First make sure your configuration menu is visible. The top-left corner of the ProcessMining container sometimes shows a hamburger menu  Click this hamburger menu because we need to access the configuration menu. You should see a menu bar on the left side of the ProcessMining container looking somewhat like the image shown to the right.

With this set of icons/menu options we will configure the ProcessMining extension.

Use the “PATHS DATA” tab to connect to your datasheet and configure the behaviour and look-and-feel of all paths in you Process Mining extension.

Select Paths Data

This is where you select the sheet that holds your data, also referred to as the ‘datasheet’

Once you selected the ‘datasheet’ the rest of the fields will be populated with all the available fields of that sheet. So next option to configure is the “Activity” field.

Activity

The activities in your process are what will be shown on as nodes in your process map, they are the events of the process. This is where you will be able to see where the real bottlenecks are in your organisation, which events are taking too much time and which are taking the happy paths. ProcessMining will clearly show which activities need more attention than others.

Case Identifier

Choose which column holds the unique identifier of the cases flowing through your process. Each case will have a complete flow of the activities and time taken. Use dashboard actions for example to link to a sheet on your dashboard to see the flow of each individual case within the extension.

Start Time

Select the Date/Timestamp from your ‘datasheet’ that represents the start of an activity. As stated before the best way to store your dates is in a discrete “exact date” value.

End time (optional)

Select the field that contains the end Date/Timestamp on your ‘datasheet’. This field is not a required field to operate the ProcessMining extension but enables you to calculate the ‘Waiting time’ it takes for cases to move from one activity to another (so the time between the end of axtivity X and the start of activity X+1)

Position of Time KPI

If only one date time stamp is specified, this option allows for the configuration of the time KPI. By default, time between start dates will be displayed on the activities. But by checking this box, they will be displayed on the paths instead.

Custom Measure

When the default KPI’s / measures in the ProcessMining extension are not sufficient for your use case you can also add an extra custom measure to your ‘datasheet’ and select it in this field to be part of your ProcessMining analytics.

Reverse Data

If the sorting in your ‘datasheet’ is not as expected but it seems to be in reversed order, this switch can easily help you out in reversing the order of your activities.

Pagination

Enabling pagination will set your data to load in chunks (or “pages). This enables use-cases with large numbers of rows.

Limit Data

When configuring or changing the extension, limiting data will improve performance. This allows users to quickly have their changes reflected in the visualisation.

Have the ability to format your paths to give yourself the most effective visualisation. Color, sizing and styles can all be changed here.

Path type

Change the style of your path using the dropdown menu. We all like choices so go ahead and choose which best suits your use case.

Adjust Path width

Check the tick box to allow the width of the path be based on the value of the selected KPI. This can give you extra value on your viz by seeing at a glance the value of each link. A path will be bigger when the KPI has a larger value.

Max line width

Use the scroll bar to change the maximum width of the line. It may not be of benefit to have huge lines throughout your process, keeping this low may give a nicer overall look, setting the value higher will differentiate more between lower and higher KPI values.

Path color

Click on the color gradient to change the color of the path shown in your viz. Once popped up click again to see the list of colors available for you. Select the color shades of your liking.

Match Path color

Before deciding on your font color, have the option to match the color of the path you’ve chosen. The Path labels will be shown in the same color as the path will be shown. So coloring of the labels will be depending on the KPI values for the links

Font Color

Pick your font color using the color palette as well as RGB, HEX and HSL. This will set a fixed font color for the labels on Paths in your Process Map.

Path Style

Allow the algorithm to choose the path style by selecting best fit or choose from the other options below. You can choose to have straight lines, curved lines, polygon lines etc. By default the style is set to “Best Fit”

Arrow Types

Choose from the selection of arrow types and use the slider to change the arrows size. Nice and simple.

Here you can change the look and feel of activities in your ProcessMining visualisation. Format your activities to give the best looks to your dashboard with respect to yout use case.

Activity Shape

Change the shape of your activities in your process map. Select a shape you like to visualize the nodes in the process map. By default the ‘Record box’ is used. This is the most common activity shape used in Process Maps. Here you have the option to have rounded edges turned on or off on your nodes. Change the whole look of your viz by having the nodes outlined only. This will trigger below the option to change the style of your outlines.

Activity Color

Click on the gradient to choose from the selection of colors for your viz. This may effect the visibility of the font color to be sure check the ‘Auto adjust font color based on color scale’ for the extension work out the best color balance. This option sets the default color of activities in your Process Map. In a later step we tell you how to set the color of activities that reside in your Happy Path. The ‘outliers’ of your happy path will have the default color that is set with this option. So if you plan on using and visualizeing a happy path, it might be wise to select a signaling color as default color…

Use absolute values for coloring nodes

The values will round to the nearest absolute node and then be colored by that value. It often results in wider color variations in your process map, just try and see what best fits your use case.

Font size

Use the scroller to change the font size within your activities.

Here is where you can format the nodes at the start and end of your process independently to the rest of your nodes in the view. The Start and End nodes are not real nodes in your process event log but they indicate where your cases start their process. The lines from/to your Start/End nodes indicate the real Start/End points that are present in your data.

Have full formatting to edit these nodes including editing the text within the nodes under ‘label’

The configuration menu speaks for it self. You can enable/disable the StartEnd nodes, set line styles, coloring and labels. It is up to you if and how you visualize these nodes on your Process canvas.

ProcessMining allows you to configure a “Happy Path”. The Happy Path is not automatically calculated, it is not the most common nor the most frequent occurring path in your dataset but can it be configured as the (imaginary) ideal path you want to be followed for the process. Outliers can be spotted immediately and action can be taken!

• The Happy Path is drawn as a vertical straight line from the Start node to the End node
• The Happy path activities and paths can be represented in a different color

Hit “Enable happy path” to have the opportunity to show the Happy Path in your Process Map. Click the ‘Visible when filter sheet(s)’ this will show the activities of the Happy Path in the visualisation even when your filtered data does not contain the Happy Path activities. The unused “Happy Path” nodes will be greyed out but still be visible. When false all non-connected nodes will be filtered from the tree.