Analyzing Process Maps

Once you have imported a data set (see Import), you are directly taken to your first process map in Disco’s analysis view. The analysis view is symbolized by the magnifying glass symbol shown in Figure 1.

_images/details-icon.png

Figure 1: Analyze symbol in Disco.

For each data set, there are three alternative analysis views as shown in Figure 2: Map, Statistics, and Cases. They show different aspects of the same, underlying data, and you can switch between them by selecting the corresponding tab. If you have applied a filter (see Filtering), then all three analysis views show you their perspective on the filtered data set.

You can use the drop-down list at the top to quickly move back and forth between your data sets (see Switching Between Data Sets in the Analysis View). Refer to Navigating From the Project View to the Analysis Screens to learn how you can return from your project overview to your analysis view.

_images/AnalysisViews.png

Figure 2: The three analysis views in Disco: Map, Statistics, and Cases.

Most of your time using Disco will be spent in the three analysis views, because that’s where you discover how your process has been running, where you inspect statistics and KPIs, and where you can track down individual cases.

This chapter explains in detail how the map view works. Refer to Analyzing Statistics and Analyzing Cases to learn how to use the statistics and cases views.

The Map View

After you have imported your event log, the first thing you usually want to see in a process mining tool is what your process actually looks like. Therefore, Disco brings you right into the Map view (Figure 3).

_images/Map-View.png

Figure 3: The Map view in Disco.

In the Map view, you see a process map that visualizes the actual flow of your process based on the imported log data. The Map view contains the following elements:

Canvas with process map (1)
The main area is reserved to visualize your process map. Refer to How To Read the Process Map for further details on how to read the process maps in Disco.
Zoom slider (2)

You can zoom in and out in your process map in two different ways:

  • The zoom slider (2) gives you an explicit control to make the process map larger and smaller.
  • Alternatively, you can simply use your mouse wheel to zoom in and out.

To move the currently displayed area of your process map around you can either use the vertical and horizontal scroll bars or click and hold your mouse while dragging the process map.

Map detail controls (3)
Because real-life processes can become quite complex and confusing when every detail and all exceptional process flows are shown, Disco gives you a quick and easy way to make the process map simpler and only show you the most important flows. Refer to Adjusting the Level of Detail in Your Process Map to learn how you can adjust the level of detail of your process maps.
Search field (4)
The search field allows you to find a specific activity also in large process maps. Read Searching Activities in Your Process Map to see how the search field works.
Process map visualization options (5)
In addition to showing you in which order the activities in your process have been performed (the actual process flows), you can also enhance your process maps with several process metrics. Refer to Frequency Metrics, Performance Metrics, and Combining Metrics for further details on the process map metrics that are available in Disco.
Animation (6)
Animation can be very useful to communicate analysis results to process owners or other people who are no process analysis experts. By showing how the cases in the data set move through the process (at their relative, actual speed), the process will be literally “brought to life”. Refer to the Process Animation section to learn more about animation.
Filtering (7)
The log filter controls for the current data set can be accessed from each of the analysis views. Filters are really important to drill into specific aspects of your process and to focus your analysis. Read Filtering for detailed information on how filtering works in Disco.
Copy, Remove, and Export data set (8)
Data sets can be copied, deleted, and exported right from the current analysis view. Read Copying Data Sets for further details on copying and deleting data sets. Process maps can be exported, for example, as a PDF file. The export functionality of Disco is explained in detail in the Export reference in Export.

How To Read the Process Map

The process map is the most important analysis result in Disco. It shows you how your process has actually been executed. The process flows that you see in the Map view are automatically reconstructed (“discovered”) based on the sequence and timing of the activities in your imported event log data. So, without further knowledge about the process, or any pre-existing process model, you obtain an objective picture of the real process.

_images/Map-Start.png

Figure 4: Start of the purchasing process demo example.

The discovered process is visualized in a simple and intuitive way: The start of the process is illustrated by the triangle symbol at the top of the process map (see Figure 4). Similarly, the end of the process is illustrated by the stop symbol (see Figure 5). Activities are represented by boxes and the process flow between two activities is visualized by an arrow. Dashed arrows point to activities that occurred at the very beginning or at the very end of the process. By default, the absolute frequencies are displayed in the numbers at the arcs and in the activities (see Frequency Metrics and Performance Metrics for how to change the metrics that are displayed in the process map). The thickness of the arrows and the coloring of the activities visually support these numbers.

_images/Map-End.png

Figure 5: End of the purchasing process demo example.

For example:

  • In Figure 4 we can see that there are 608 cases (different instances of the purchasing process) in the data set that all start with the activity Create Purchase Requisition.
  • Afterwards, the process splits into two alternative paths: In 374 cases the activity Analyze Purchase Requisition was performed after Create Purchase Requisition. The other 234 cases perform activity Create Request for Quotation Requester instead. Because the path where 374 cases have “travelled through” indicates the main flow in this part of the process, it is visualized by a thicker arrow.
  • In total, activity Analyze Request for Quotation is the one that is executed most often (in total 1107 times)—almost twice as much as we have cases in the data set! This comes from the dominant loop with activity Amend Request for Quotation Requester (see Figure 4). Repeatedly, purchase orders are amended and need to be re-analyzed, which is of course very inefficient and from a process improvement perspective we would need to find out what is going on. Perhaps people don’t know what they are allowed to purchase, and we might resolve the problem by updating our purchasing guidelines or providing additional training.
  • Figure 5 shows another fragment from the end of the purchasing process. We can see that 413 purchase orders are completed and end with Pay invoice. Some others are stopped earlier in the process (see other dashed arrow). In fact, in Figure 4 we can see that some are stopped after activity Analyze Request for Quotation (see dashed arrow at the bottom of the shown fragment).

Adjusting the Level of Detail in Your Process Map

Real-life processes can often become very complex. Therefore, Disco allows you to interactively adjust the level of detail that you want to see. There are two slider controls that you can use to modify the level of detail that is shown in your process map:

Activities
The Activities slider influences the number of activities shown in your process map, ranging from only the most frequent activities up to all including the least frequent activities.
Paths
The Paths slider determines how many paths are shown in your process map, ranging from only the most dominant process flows among the shown activities up to all (even rare) connections between the activities.
_images/Activity-slider-low.png

Figure 6: The Activities slider influences the number of activities shown in your process map. It ranges from showing only the activities from your most frequent process flow at the lowest point (0%) up to showing all activities that have ever occurred (100%).

In Figure 6 you can see how the purchasing example process from the previous section looks like when the Activities slider is set to the lowest point. Only those activities that occur in the most frequent process variant are shown.

When you move the Activities slider up, increasingly also less-frequent activities are included, up to showing all activities that have ever occurred (even if they just occurred once or twice). For example, in Figure 3 and in Figure 4 you see a process fragment with 100% of the activities shown: Less frequent activities such as Amend Purchase Requisition, which occurred only 11 times, are shown as well.

_images/Path-slider-high.png

Figure 7: The Paths slider can be used to show only the most dominant paths in your process map (0% slider position) up to all connections between activities that have occurred (100%).

The Paths slider works based on the activities that are currently revealed by the Activities slider (e.g., all or just a subset). If the Paths slider is set to the lowest point, then only the most dominant connections between these activities are shown. This way, Disco makes sure that all your activities are always connected and avoids getting “dangling” process fragments that cannot be put in context with the remaining activities even if you look at a simplified process map. For example, in Figure 3 and in Figure 4 you see the purchasing process with all activities (Activities slider pulled up to 100%) and just the most dominant connections between them (Paths slider set to the lowest point).

In Figure 4 you can see that not all the details of the process are revealed yet because activity Amend Purchase Requisition has been performed 11 times directly after activity Analyze Purchase Requisition (see incoming path with number 11), but only 8 times the process has returned to Analyze Purchase Requisition afterwards (see outgoing path with number 8). The numbers don’t match up yet. Where did the process go in the other 3 cases? This can be revealed by pulling the Paths slider up to 100% as shown in Figure 7. We can see that 3 times the process went directly from Amend Purchase Requisition to activity Create Request for Quotation Requester.

To determine the right level of detail for your own process, it is recommended to start with the slider position that Disco determines automatically when it creates the first process map for you. Then, start by pulling up the Activities slider until you see all the activities that you want to see. Most of the times, you will be able to reveal 100% of the activities and still get a readable process map. Only then start moving the Paths slider up as much as possible. Stop when the process map becomes too complicated to inspect it properly. This way, you can get reasonably complete and still readable process visualizations for almost any process, no matter how complex it is.

If you want to repeat your analysis on a new data set, or compare the process across multiple data sets, you often want to keep the slider positions constant. However, moving the sliders to hit the exact percentage point that you want to see can be cumbersome. Therefore, you can explicitly set your percentage for the Activities and Path sliders by clicking on one of the % labels below the sliders (see Figure 8).

_images/Explicit-Sliders.png

Figure 8: By clicking on the label below the sliders, you can set a fixed percentage for the Activity and Paths detail sliders.

The simplification sliders are great, because they can deal with very complex processes. Furthermore, the full scope of the process (all activities) can often be preserved, which makes sure that also low-frequent activities can be investigated. At the same time, for most unfiltered data sets it is rarely possible to reveal all the paths, which can lead to confusion when you share or present your process maps outside of Disco, because the numbers don’t add up (yet).

If you find this to be a problem, note that you can also simplify your process map by focusing on the main variants. Refer to the Variation Filter to learn how to filter based on process variants and to the article series Managing Complexity in Process Mining [1] for further simplification strategies.

Searching Activities in Your Process Map

When you have process maps that contain many different activities, it is sometimes difficult to find a specific activity that you are interested in. In such situations, you can use the search field in the upper right corner of the process map.

_images/Search-Map.png

Figure 9: When you start typing in the search field, Disco interactively highlights all activities that contain the current search term for you.

_images/Search-Map2.png

Figure 10: When you type multiple words, you can narrow down your search to activities that contain all of these words.

Searching works as follows:

  • You simply start typing in the search field as shown in (1) in Figure 9. While you type, Disco highlights all activities that match the searched term (see bold red arrows in Figure 9) and automatically zooms into the process map to bring all matched activities into focus.
  • You can narrow down your search by typing multiple words as shown in Figure 10.
  • If you want to go back, you can simply delete the word that you typed in the search field by pressing the Delete key.
  • If you want to stay in the focused view but get rid of the red arrow markers, press the little x button in the right corner of the search field (see in (2) in Figure 9). This will remove the bold red arrows but keep the zoom level and position.

Frequency Metrics

In the default view, the process map is displayed with absolute frequencies. This means that the numbers in the activities and at the paths indicate how many times the activity was performed in total, or how often that particular path has been “travelled”, respectively, throughout the whole process. However, you can change the visualization and display also other metrics right in your process map.

The frequency metrics show you which parts of your process have been executed most often. You can change between them like shown in Figure 11.

  • Absolute frequency. This is the default view based on total frequencies as explained in the beginning of this chapter: The total number of times that a particular process step was performed, or a particular process path has been followed, is added up over the whole data set (see also Figure 11).

  • Case frequency. If you have repetitions in your process, it can be helpful to ignore them for a minute and just view relative numbers of how many cases passed through which activities and along which path (regardless of whether they came by there just once or multiple times).

    An example of the Case frequency view is shown in Figure 12, where—in contrast to the absolute view in Figure 11—it now becomes visible that out of 608 cases 231 went through the loop with activity Amend Request for Quotation Requester.

  • Max. repetitions. Displays the maximum number of repetitions within a case. The Max. repetitions option does the opposite of the case frequency: Instead of ignoring repetitions within the same case, it exactly focuses on up to how often the process passed through a particular activity or path for the same case.

    For example, in Figure 13 we can see that the loop with activity Amend Request for Quotation Requester was performed up to 12 times within a single case.

_images/Frequency-1.png

Figure 11: Within the Frequency or Performance view, you can select the metric that you want to display from the drop-down list as shown here.

_images/Frequency-2.png

Figure 12: Use the Case frequency option to ignore repetitions and just see relative numbers for how many cases passed through which activities and along which paths.

_images/Frequency-3.png

Figure 13: The Max. repetitions option lets you see up to how many repetitions occurred in your process within the same case.

Performance Metrics

When you have obtained a good understanding about the actual process flow, you often want to know more about the time that is spent in the different parts of your process. This is what the Performance metrics are for. You can change to the Performance view by selecting the Performance tab, which brings up the performance metrics and visualization legend (see Figure 14).

  • Total duration. The default metric that is displayed when you get into the performance view is the total duration. It shows the accumulated durations (summed up over all cases) for the execution of each activity and for the delays on each path.

Because the frequencies of the activities and paths are included in this cumulative view, it allows you to quickly spot the high-impact areas in your process: For example, it can have much more impact on your overall throughput time if you can speed up a frequently performed part of your process by just 5 minutes compared to reducing the time spent in a rarely used process path by 1 day.

An example of the performance view with the Total duration option is shown in Figure 14. There, we have filtered the data set to only show the process map for those cases that last longer than 70 days (refer to the Performance Filter to find out how to do this), and now we want to know where all this time is lost in the process. Clearly, the biggest impact area is the delay between the activity Amend Request for Quotation Requester and Analyze Request for Quotation (see thick red arrow with the displayed cumulative time of 11 years). If you are confused by the big numbers (the cumulative delays easily add up to months or years), read the section on Combining Metrics to learn how the total duration can be combined with other metrics.

  • Min. duration. The smallest execution times and delays that were measured.

    The minimum duration statistics can be useful when you want to know how fast a particular process step or process transition can be. You can then filter for cases, where particular parts of the process were really fast (see Follower Filter) to investigate these cases further and see whether there are best practices that can be promoted from them to the overall process.

    On the other hand, the minimum duration can also highlight problems. If, for example, an authorization activity from a manager that normally takes 20 minutes has a minimum duration of only 10 milliseconds, you know that you are either dealing with a suspicious situation, such as that there was not actually a thorough review of the application done by the manager, or that there are data quality problems in your data.

  • Median duration. The median time spent within and between activities can be displayed. In many situations, the median (also known as the 50th percentile) gives you a much better idea of the typical performance characteristics of a process than the arithmetic mean, especially for data sets that contain extreme outliers.

While the median is very useful for analysis, it is quite demanding to determine, both in terms of computing power and regarding memory requirements. Disco features an optimized median computation algorithm that can compute the precise median all over Disco with a significantly reduced memory footprint. When you have a huge or complex data set, and Disco runs low on available memory, it will automatically transition to an advanced and more memory-efficient calculation method that can can estimate the value of the median with a very low error margin for selected medians. You can distinguish precise medians from medians that have been transitioned to the more memory-efficient calculation method by the tilde prefix. For example, in Figure 15, the path with the “~ 142 milliseconds” median duration has been estimated, while the other paths (with “3.9 d” and “71.1 mins”) are precise.

Unless you are working with very large data sets, you will probably never see an estimated median in Disco. And even when you do, in all likelihood the estimated median will differ only very slightly from the precise median, or not at all. And for those rare situations when you absolutely do require total precision of all medians in a huge data set, you can simply increase the memory available to Disco in The Control Center. The median calculation system in Disco provides the best of both worlds. Wherever possible, you get an absolutely precise median with the minimum memory footprint and best system performance. Whenever that is not possible, Disco automatically reduces the precision for those measurement points where it makes the least difference. In that way, you will get nearly precise medians also for very large data sets.

  • Mean duration. Alternatively, the average time (arithmetic mean) spent within and between activities can be displayed. The mean has the advantage that it is more widely know than the median (see above). So, it may be easier for people without statistical knowledge to understand.

    For example, in Figure 16 one can see that while the execution of activity Amend Request for Quotation Requester takes on average only 9.8 minutes, it creates a delay of, on average, 14.9 days afterwards.

  • Max. duration. The largest execution times and delays that were measured.

    The maximum duration statistics show you the worst cases in terms of how long a particular process step or path has taken in the process. One use case for the maximum durations is also to distinguish automated process steps from steps that are performed by humans. If a process step is automatically executed, there is of no or almost no delay. So, these steps or paths will show up with a very small, or instant, maximum delay.

_images/Performance-1.png

Figure 14: The Total duration option lets you see the high impact areas for delays in your process by showing the cumulative times (added up over all cases) for each path and activity.

_images/Performance-Median.png

Figure 15: When you use the Median durations in your process map, you may sometimes see the tilde prefix if you are working with very large data sets. By pre-fixing the median value with a tilde, Disco lets you know that this median is not a precise calculation but was estimated with a very low error margin.

_images/Performance-2.png

Figure 16: The Mean duration option displays the average execution times for each activity and the average idle times on each path.

Note

The execution times for activities can only be displayed if you have both start and a completion timestamps in your data set. If you only have one timestamp in your event log, then you can still analyze the time between activities, but the activity durations will be displayed as instant.

Refer to the import reference in Including Multiple Timestamp Columns for further information on start and completion timestamps, and how they can be included.

Combining Metrics

Sometimes you just want to see all these metrics at one glance and not switch back and forth. This can be achieved by simply clicking on an activity or path as shown in Figure 17 and works regardless of in which metrics visualization view you currently are.

When you want to remove the overview badge again, just click somewhere at the background of your process map (outside or next to an activity or path).

_images/Popup-Metrics.png

Figure 17: When you click on an activity or a path in your process map, you can see all metrics at one glance in an overview badge.

In addition, you can also display two different metrics in the process map in the following way. You can add a secondary metrics to your process map visualization, by clicking on the Add secondary button below the perspective legend on the bottom right (see Figure 18). The primary metric will still determine the visualization (colors, shades, thickness of the arrows) of your map to guide your attention. But now, the labels of both activities and paths will also feature a label detailing the secondary metric in smaller font below (for the paths paths) and in brackets next to (for the activities) the primary metric value.

All Frequency Metrics and Performance Metrics can be combined in this way.

_images/Secondary-Metric.png

Figure 18: To display an additional metric in your process map, click on the Add secondary button below the legend.

One situation, where combining two metrics in your process map is when you perform a bottleneck analysis. For example, in Figure 16 we were inspecting the average delays in the process. However, neither the mean nor the median metrics take the frequency into account. This means that we may see big delays in a process step, or a process path, that was performed just a few times. Improving this part of the process will not have much of a big impact on the overall process.

If you use the total duration metric as shown in Figure 14, you do see the places, where the biggest delays accumulate overall. However, the numbers are less meaningful to understand how long these delays are for a typical case, because they are added up for the whole data set.

To keep the focus on your most significant bottlenecks and easily see the typical delay for each process step and path, you can keep the Total duration as your primary metric and add either the Mean duration or the Median duration as a secondary metric. In Figure 19 you can see this combination for total duration and median duration. The bottleneck can be clearly identified (see the big, red arrow from Amend Request for Quotation Requester to Analyze Request for Quotation, which cumulates to 11 years of total duration), while you can read off the median duration (11 days) as well.

_images/Bottleneck-Analysis.png

Figure 19: To keep your focus on the high-impact areas while still being able to see the average delays, you can combine the Total duration (primary metric) with a secondary metric such as the Mean duration.

Filtering Paths from the Process Map

The overview badge shown in Figure 17 also allows you to quickly filter your data set for cases that follow a particular path, or that execute a particular activity.

For example, in Figure 20 you can see the end of the purchasing process from before. Clearly, the main process flow runs through the activity sequence Send invoice -> Release Supplier’s invoice -> Authorize Supplier’s invoice payment (potentially settling a dispute with the supplier in between). In fact, the activity Authorize Supplier’s invoice payment is a mandatory process step that was put in place to avoid fraud.

However, one can see that 10 cases move directly from Send invoice to Authorize Supplier’s invoice payment, skipping the required Release Supplier’s invoice step. This is a compliance issue. As a process analyst we would want to find out which cases have bypassed the prescribed process step to find out what happened there, and to prevent this in the future.

_images/FilteringPath-1.png

Figure 20: 10 cases bypass a mandatory authorization step at the end of the purchasing process. As a next step, we want to drill down to find which 10 cases these were.

Drilling down into specific cases that follow a specific path is easy: You can simply click on the corresponding path to bring up the overview badge as shown in Figure 21.

_images/FilteringPath-2.png

Figure 21: Step 1: Click on the corresponding path to bring up the overview badge. Click the Filter... button to add a filter for this specific path.

Once you press Filter this path... a pre-configured Follower Filter is automatically added to the filter stack (Figure 22). After you have applied the filter you get back to the process map, filtered down just for these 10 cases (Figure 23).

_images/FilteringPath-3.png

Figure 22: Step 2: Disco automatically adds and pre-configures a Follower Filter. Just press Apply filter to activate the filter for your data set.

However, in this situation the process map is not particularly interesting. Instead, we want to see which cases followed the non-compliant path. So, you can change to the Cases view (see Analyzing Cases) as shown in Figure 24 to see all the details about the 10 cases that followed the forbidden path.

_images/FilteringPath-3b.png

Figure 23: Step 3: Change to the Cases view (Analyzing Cases) to see detailed information about these 10 cases rather than the process map.

_images/FilteringPath-4.png

Figure 24: Step 4: The Cases view now shows a list of all cases in the filtered data set. Here, case 812 has been selected and we can inspect the history with all its details.

Filtering Activities from the Process Map

In Filtering Paths from the Process Map, a path was filtered and a corresponding Follower Filter has been added automatically (Follower Filter) from the process map. In the same way, an activity can be filtered by clicking on the activity and pressing Filter this activity... as shown in Figure 25.

_images/FilterActivity.png

Figure 25: Clicking on any activity brings up the Filter this activity... shortcut to add a pre-configured Attribute Filter in Mandatory mode that only keeps those cases that pass through this activity.

This will add a pre-configured Attribute Filter (see Figure 26) that only needs to be applied to narrow down the data set to those cases that perform the activity (here this will be the 80 cases that performed activity Settle dispute with supplier).

_images/FilterActivity-2.png

Figure 26: Disco automatically adds and pre-configures an Attribute Filter in the right filter mode. Just press Apply filter to activate the filter for your data set.

Filtering Start and End Activities from the Process Map

In addition to Filtering Paths from the Process Map and to Filtering Activities from the Process Map, you can also filter start and end points directly from the process map. To do this, simply click on the dashed line leading to the start point (or the dashed line leading from the end point) that you want to filter.

For example, in Figure 27 we have clicked on the dashed line leading from the activity Pay invoice to the end point.

_images/FilterEndpoints-1.png

Figure 27: Clicking on the dashed line brings up the Filter for this end activity... (or Filter for this start activity..., respectively) shortcut to add a pre-configured Endpoints Filter that only preserves cases that end (or start) with this particular activity.

After pressing the Filter for this end activity... button, a pre-configured Endpoints Filter will be automatically added to your filter stack. All you have to do to narrow down your data set on the 413 cases that end with the activity Pay invoice, is to apply the filter (see Figure 28). This works in the same way for start activities in your process.

_images/FilterEndpoints-2.png

Figure 28: Disco automatically adds and pre-configures an Endpoints Filter in the right filter mode. Just press Apply filter to activate the filter for your data set.

Process Animation

The animation is a way to visualize the process flow over time right in the discovered process map (a bit like showing a “movie” of your process). Animation should not be confused with simulation. Rather than simulating a process, the real events from the log are replayed in the discovered process map as they took place (based on the timestamps in your data set).

The animation can be very useful to communicate analysis results to process owners, or other people who are no process analysis experts. By showing how the cases in the data set move through the process (at their relative, actual speed), the process will be literally “brought to life”.

For example, imagine that we want to visualize the bottleneck that we discovered in the purchasing process, when we analyzed the performance metrics in the process map for cases that take longer than 70 days in Combining Metrics. To start the animation, simply press the little button with the Play symbol at the bottom of the Map view, see also (6) in Figure 3. This will bring up the animation view as shown in Figure 29:

_images/Animation.png

Figure 29: Animation view in Disco. Start the animation by pressing the Play button (1) and observe how your cases flow through the discovered process map over time.

Play button (1)

Start the animation by pressing the Play button in the lower left corner. You will see small “bubbles” (so-called tokens) that start moving through the process map. Each token represents one case and moves through the process at the relative, actual speed of the corresponding case in your event log.

  • Between activities, the cases are visualized in bright yellow color with a red border to make it easy to spot unnecessary delays (like, for example, there is a clearly visible queue of cases in the back loop from activity Amend Request for Quotation Requester -> Analyze Request for Quotation in Figure 29).
  • When an activity is performed, the active activity is highlighted in blue. It slowly fades back to grey after the activity has stopped to let you observe activity sequence patterns as they unfold.

Cases that are currently performing an activity are visualized in a light blue color inside the activity box (for example, in activity Send Request for Quotation to Supplier in Figure 29 there is currently one case performing this activity).

When there are two activities that are performed in parallel for the same case, then this case is visualized by two tokens for that time.

The animation is performed in chronological order based on the timestamps in your data set. It starts with the cases that were active at the beginning and ends with the latest activities in your log.

Progress indicator (2)

The progress indicator shows you how much of your overall log timeline has already been replayed by the animation. The replayed part is highlighted in blue and the needle indicates the current replay position. To move around in the timeline simply drag the needle to the left or right.

In addition to the position of the animation in your replay, the timeline visualization also gives you a sense of how much activity occurred in your process over time (the thicker the more activity). This helps you, for example, to jump right into the more “busy” periods of your process with the animation.

Current replay time (3)
The current replay position is displayed in this date and time window. For example, in Figure 29 the animation shows the state of the purchasing process on Friday 10 June 2011 at 20:14.
Speed control (4)
To increase or decrease the speed of the animation, you can move this speed slider to the left (slower) or to the right (faster).
Export animation (5)
To export your animation as a movie file, press the export button and choose the place on your computer where the animation movie should be stored in the upcoming file dialog. Refer to Exporting Animation Movies to lear more about exporting animations.
Zoom control (6)
You can zoom in and out in the animated process map by using this zoom slider or, alternatively, use your mouse wheel to zoom in and out (works in the same way as in the regular Map view). To move the currently displayed area of your process map around you can either use the vertical and horizontal scroll bars or click and hold your mouse while dragging the process map.
Full screen mode (7)
For presentations, it can be useful to focus only on the animation and use as much of your screen as possible to display the process flow. For this, you can press the full screen mode button in the upper right corner. When you want to return from the full screen mode, simply press the same button in the upper right corner again.
Return to Map view (8)
To return to the regular Map view of your process, press the Exit button in the upper left corner.

One primary use case for animations is when you want to intuitively spot and highlight bottlenecks in the process. In such bottleneck situations, you have a lot of cases that accumulate in certain paths. However, if they would just lie on top of each other, you would not be able to recognize the bottleneck.

Therefore, when many cases pile up on a certain arc and are causing congestion, Disco groups these cases into larger “bubbles” (see Figure 30).

_images/animation-congestion.gif

Figure 30: If many cases are piled on top of each other, bigger bubbles are created to highlight the bottleneck in your process.

This allows you to immediately spot the worst bottlenecks and focus your improvement efforts. It is also a great visual effect, allowing you to get a much more intuitive feel for the dynamics of cases moving through your process.

Synchronized Animation

Normally, the animation replays your process chronologically. This means that the process is visualized based on the actual timestamps for each case over the whole timeframe of your data set. This is great to look at the process how it actually happened, to spot process changes over time, and to identify busier and less busy periods in your data set (for example, discovering seasonal patterns).

In contrast, the synchronized animation starts to replay all cases in your data at the same time. So, the replay of all cases is shown relative to each other. This allows you analyze at what time into the case execution the hot spots and bottlenecks in your process are most prominent, and to compare your process performance over the set of cases in your data.

In Figure 31, you can see a synchronized animation for all purchase orders that are running longer than 70 days. Instead of the date and time, the replay time window now shows you how long all cases that are currently visible have been already running. Also the progress indicator now spans the range of throughput times in your data set rather than the timeframe of the data set. For example, in Figure 31 you can see that at the current replay state 80 days, 17 hours, 40 minutes, and 45 seconds have passed. For all the cases that are still in progress you can see where in the process they are currently waiting.

_images/AnimationSynchronized.png

Figure 31: With the synchronized animation you can watch the flow of all cases relative to each other. The replay time now indicates how long the cases have been running and in the animation you can see where in the process each case is at that stage.

You can choose between regular and synchronized animation by right-clicking the animation button in Disco’s process map view (see screenshot in Figure 32).

_images/StartSynchronizedAnimation.png

Figure 32: Right-click the animation button and choose the Synchronize case start times option.

“I cannot find the animation button”

If you do not see the Animation button at the bottom of your process map, then you have imported a data set without timestamps.

If you have timestamps available in your data set, you can import the file again and make sure that at least one column is configured as a timestamp. This will then enable the animation option in Disco.

“Disco tells me that my data set is too large for the animation”

While Disco implements the live animation using efficient and performance-optimized graphics functions, it still remains a resource-intensive part of the application. Therefore, it can happen that a larger data set (for example, a data set with many millions of events) is too big to be animated completely. If your data set is too large, Disco will give you the following message (see Figure 33) and create the animation based on a subset of your data.

_images/AnimationError.png

Figure 33: It can happen that your log is too big to create a full animation. Disco will tell you how many cases it could use and create the animation based on that subset if you click OK. If you want to animate the full data set, you can then either allow Disco to use more main memory or use a filter to created a smaller data set.

You have three options to deal with this situation:

Option 1: Use the animation based on the subset that Disco has created
You can simply press OK and use the animation based on the subset that Disco has created. In the screen in Figure 33 you can see that Disco gives you an indication of how many % of the cases could be included in the animation. These are the cases for the first xx % Case IDs that appear in your source data file.
Option 2: Increase your main memory

Normally, your analyses in Disco are not limited by the memory in your computer, because data that does not fit in the main memory is written back to the hard disk and read from there transparently when you filter, create process maps, etc. However, the animation is the one place in Disco that is currently limited by the memory that you have available.

Per default Disco only uses 1 or 2 GB of RAM (depending on the operating system). If your computer has more RAM, you can use the ‘Optimize memory’ button to increase the amount of memory that Disco can use and, therefore, increase the volume of data that can be animated. Refer to Increasing Your Main Memory to see how to do that.

Option 3: Use a filter to reduce your data set
If you want more control over which subset of your data is animated, you can use a filter to reduce your data set and animate only a part of it. For example, try using the Timeframe Filter to split your data set into two parts, one for the first half and one for the second half, to animate them after each other.

Footnotes

[1]The article series is available at the following URL http://fluxicon.com/blog/2015/03/managing-complexity-in-process-mining-quick-simplification-methods/.