Visualise a Business Process in PowerBI

Business Process Analysis is a methodology for the analysis of a business with a view to understanding the processes and improving the efficiency and effectiveness of its operations. In this article we explore the concept of visualising this idea, using Power BI, from transforming a dataset into an event log to incorporating a custom visual in the form of a sankey diagram to provide a clear and succinct overview of the processes captured within the data.

Akin to the concept of Process Mining, utilising an analytical method such as this allows a user to quickly and immediately obtain an overview of the process each line of their data is taking, be it P2P, O2C or otherwise, the data needs to have captured several ‘events’ with a timestamp (i.e. a date field) alongside key information relating to those events (such as user detail or invoice information).

Below, we give an overview of a recent case where this idea was implemented for a P2P process:

1. First we extracted  key line-level fields from the dataset for grouping and transformed the date fields so they were appended on top of each other, rather than horizontally at row-level.

2. Next, we manipulated this further so that, for each line, date information from the line below was added alongside to achieve a Source – Destination combination for every row, iterating down the list for each key field grouping. Dates which occurred on the same day within a key were defaulted to be assumed as occurring in the ‘standard’ order of process. Additionally, creating an Event Number that increased consecutively for each line within a grouping of key fields allowed us to automate the Sankey diagram applying the correct ordering and distribution to the various processes.

3. Finally, we applied this manipulated data to the field inputs for the Sankey diagram custom visual. From here, we incorporated filters by linking the key fields back to the original dataset to identify exceptions to the accepted process order.

To conclude, utilising an approach such as this within dashboards provides the user with several key advantages:

• You get a complete overview of your dataset, with a focus on dates, in one succinct visual.
• You can visually identify potential process control failures, drill down into the detail and take action on high risk processes.
• You can monitor progress towards better conformance of processes.
• You can identify ‘bottlenecks’ within the process where events are taking longer than expected by calculating a timestamp between each pair of events. Plus, you can drill down into the detail to identify further trends using additional key information from the original dataset.

Author:- Samantha Kay, DataConsulting