Process data analysis

In Signavio Process Intelligence, an analysis view that explains a specific process insight is called an investigation. After you have uploaded your data, you can create a new investigation by clicking New investigation and entering a name for your analysis. You can create multiple investigations per process. Using this, you can clearly distinguish between different business challenges.


Filters help you create a view on an aspect of your process data that you’re most interested in. You can apply filters on a widget, chapter or document level, which allows you to structure your investigation as you desire. Once done, you can then click the + sign on the right-hand side of your chapter titles to apply one or multiple filters to your process data.

After applying a filter, you can see the absolute number and the percentage of cases the filter includes in the top right corner of the investigation or chapter.

You can create multiple filters per investigation. These are then combined by a logical AND operator.

Widgets organize and visualize information about your process. For example, the process discovery widget provides insights about the sequence flows and distribution of a process’ variants (traces). To create a new widget, click the Add widget button at the bottom of your investigation, select a widget type and specify further widget configuration options (depending on the widget type), then click save.


The filters and available widget types allow you to drill down into specific process aspects and create comprehensive analyses. We recommend using a range of widgets to document exactly how you derive your analysis results. This makes it easier communicate the results later on.

Widget types

Here is an overview of all widgets currently available for use in Signavio Process Intelligence.

Activity list

The activity list widget shows all the activities in your process. This widget is divided into two sections: conforming activities and non-conforming activities. This lets you see at a glance what the activities in your process are actually doing, compared to your model. Each section gives you the number of cases that fit each category, along with the percentage.

Bar chart

Use this widget to generate a bar chart from your data. This is useful for situations such as visualizing your cases by issue type for an investigation about support times. You can choose to aggregate your data by choice, duration or currency. Aggregation functions are COUNT, SUM, AVG, MIN or MAX.

Data grouping can also be done by choice, duration or currency. You can choose the orientation of your graph, as well as add a name. You can also add a threshold line or region indicators to your graph. To do so, first create your bar chart. Then, click the three dots in the upper right corner of the widget. Hover over “Threshold” and select none, line, two regions or three regions. From there, you can set the values you want for your thresholds and regions.

Cases table

This widget generates a table with case information you select.


Use a histogram widget to show the distribution of of duration attributes–you can see at a glance how much time cases take to complete. You can select the data aggregation source (cycle time or duration) as well as how cases are grouped. You can also add a threshold line or region indicators to your graph. To do so, first create your histogram. Then, click the three dots in the upper right corner of the widget. Hover over “Threshold” and select none, line, two regions or three regions. From there, you can set the values you want for your thresholds and regions.


You can use this widget to display diagrams from your Signavio Process Manager workspace. This is helpful if, for example, you’re investigating an issue-to-resolution process, and want to display your customer journey map in your investigation.


The distribution widget displays the number of cases that are distributed according to a selected attribute. For example, if you were investigating the amount of cancelled orders your company ends up receiving, you could use the distribution widget to see the ratio of cancelled to completed orders in your process.

Process discovery widget

Use this widget to start your investigation into your processes. It generates a process model from the event log data, visualizing the most important behavior of your process. It also has a Google Maps-like feature that lets you zoom in or out on your process model. You can choose to have your model focus on either occurrences or cycle time. The resulting model lets you focus on the most important aspects of your data by giving an overview of multiple variants at once, and lets you see what activities tend to happen the most after each other.

The numbers on the right-hand side represent how many activities and sequence flows in your process are represented in the current model. 100% activities would mean the model is showing all activities of all cases in your data set. When you view your data in the process discovery widget, the algorithm looks for activity sets with the highest frequency. These activities are the most important ones in your process. For example, if you model has 60% activities shown, it means that the other 40% are not relevant or important to what aspect of your process you are investigating.

You can increase or decrease the amount of activities and sequence flows in your model by using the sliders. The complexity of the model will be adjusted accordingly. By default, the sliders are set at the bottom, with no activity number displayed. The number of edges between activities changes according to connections between activities. Meanwhile, the thickness of the lines indicate either the amount of cases or the amount of time between events when analyzing by occurrence or cycle time.

Process variant widget

Variants are helpful to see what pathways are followed through your process. On the left-hand side of the process variant widget, your pathways are displayed as colored lines. The thickness of these lines depends on how you’re viewing the variants: by occurrence or duration. For occurrence view, the thickness of the line indicates how many selected variants are in your process. In duration view, line thickness indicates how long a pathway takes. Regardless of which view you’re using, you can click on the lines to see details of the duration and number of occurrences in your case. The line color changes depending on if the variants conform or don’t conform to your process.

Sorting can be done by criteria such as increasing or decreasing number of cases or length of duration. You can also sort by case attribute, including currency amount.

On the right-hand side of the widget is the number of variants in your process. The numbers in this table change depending on what attribute you’ve chosen. You can choose to view this attribute by number of cases, or percentage by clicking the button in the upper right corner. If you want to drill down into specific cases for a certain variant, you can change the filter. Click either the clock or the wallet icon and select a different attribute from the drop down menu.

Finally, you can also export your selected variants as a BPMN model in Signavio Process Manager. Click the ‘Open as BPMN’ link. Signavio Process Manager will open in a new tab, and display your model in the Editor. Remember to save your model before closing the tab.

Process conformance widget

Similar to the process variant widget, but the process conformance widget maps variants against a BPMN model of a process. This way, you can see what your actual process is like compared to your process model. You must first map your investigation to a BPMN model to use this widget.


Use the spreadsheet widget to create an Excel spreadsheet of your data in your investigation. The usual calculations (SUM, ADD, COUNT, MULTIPLY) apply.

Funnel diagram

The funnel diagram widget displays a visualization of traffic as it moves through your process. It is based on Sankey diagrams. To use this widget, you must first link it to a process diagram.

In this widget, the traffic patterns in your data are displayed as blue lines. Line width indicates how much traffic is following that particular pathway. Click the - or + buttons to add or remove additional pathways from the display.

You can track where customers enter and leave your process by looking at the circles at the end of each line. A circle with a triangle indicates where traffic enters your process, while a circle with a square indicates where traffic exits. Circles with numbers show the same thing, but where traffic enters and exits your process outside of the main traffic pattern. You can choose to toggle on or off the number of cases following each pathway.

Time series

The time series widget displays developments in your process over time. For example, you can view cycle time over days and weeks, the amount of cases during a certain duration, the volume of help requests, and so on. You can also add a threshold line or region indicator to the widget. To do so, first create your time series widget. Then, click the three dots in the upper right corner of the widget. Hover over “Threshold” and select none, line, two regions or three regions. From there, you can set the values you want for your thresholds and regions.


The text widget allows you to add rich text sections to your investigation. This can be helpful to, for example, explain your initial assumptions or to interpret a data chart.


The value widget aggregates case data. For example, it can be used to show the average purchase value of a subset of the cases of a process. You can also select one of two threshold display modes: two region (for differentiating between good and bad) or three region (displays a configured value next to a bullet chart). The following aggregation functions are available:

  • SUM
  • AVG
  • MIN
  • MAX

Variable importance

Similar to the bar chart widget, the variable importance widget generates a bar chart related to variables you choose. For example, if you choose to sort by currency, you can see all the variables in your process that have a strong relation to that variable.