Test what you’ve learned in week 1

Consider the following example data table.

If we want to analyse how the website is browsed by individual users, what would we need to select as our case notion?

Consider a process where papers need to be reviewed before acceptance for a journal: Reviewers need to be invited, reviews need to be obtained and summarized. Based on the reviews a decision has to be made whether a paper will be accepted or rejected. Additionally, the editor may decide to invite an additional review in case the decision is not clear. Once a decision has been made this is communicated to the authors.

If we want to analyse how the review process for a paper is handled, what would the case notion be?

Load the “Artificial - Review - Large.xes.gz” event log in ProM. This event log contains the simulated events for this process. Visualize the event log using the log dialog.

How many events are contained in this event log?

From the log dialog, what timespan is covered in the event log?

In the log dialog go to the summary tab. What is the most frequent activity?

Now visualize the event log using the dotted chart. Set the ‘color attribute’ to ‘MXML Legacy Classifier’ and set ‘Trace Sorting’ to ‘Sort on time:timestamp of first event’ (this is the 8th option from the end of the list). What can we say about the arrival rate of new cases in this process?

Change the settings of the dotted chart to show the case durations. Set the ‘X Axis Attribute’ to ‘E: time:time SinceCaseStart’ and the ‘Trace sorting’ to ‘sort on time:duration of trace.’. What can we observe on the duration of the cases?

Change the settings of the dotted chart to show when users were active. Set the ‘X Axis Attribute’ back to ‘E:time:timestamp’ and the ‘Y axis Attribute’ to ‘C: Resource’, and the ‘Trace sorting’ to ‘sort on time:timestamp of first event’. Check all answers that are correct (note: multiple answers (including none or all) might be correct!),

Select all the answers you think are correct.

Go back to the ‘Log visualizer (LogDialog)’ and switch to the summary tab. Inspect the start and end activities. For process discovery we prefer traces that have completed the full process. Do we need to filter out traces from this event log that did not complete the process yet?