Day 3 of dashboard week was Power BI day. We were challenged to create a dashboard with a focus on storytelling. We were limited in our data choice to only Victorian crime statistics from the government. Follow along as I take you through how I tackled this challenge.

After browsing the data, I chose to look at the different property that was stolen in different areas of Victoria, Australia. Within this Excel file was multiple tables but only three tables piqued my interest:

  1. Property stolen by LGA
  2. Offences at farm location by LGA
  3. Types of Property Stolen by Item Type (for farm locations)

The first challenge I had to face here was when I exported the data into Power BI, an error came up for the postcode. Due to the constraint of time, I ultimately decided that the best way to resolve this was to clean and prepare the data in Alteryx instead of Power Query. After plugging the three tables I chose to look at into Alteryx and outputting them as .csv files, I no longer had errors coming up in Power BI.

The second challenge I ran into was whether it would be appropriate to merge these tables together because each table had a different level of detail. After preparing, cleaning and exploring the data in Alteryx more, I subsequently decided that it would be best to create different charts that have limited interactivity between each other, which lead to my third challenge.

To further explore the data, I utilised my insights matrix, which you can read about in this blog post I wrote. This insights matrix helped me expand on the insights I can find from a limited number of fields in a dataset. I created an insights matrix for each of the three tables to ask questions about the relationship between two fields.

Property Stolen by LGA:

Offences at Farm Locations by LGA:

Types of Property Stolen by Item Type (for farmland):

From my the insight matrix I made for each table, I found common questions I wanted to ask which included:

  1. Is particular property types being stolen more over time, is theft happening more often, and is the value of stolen property increasing over time?
  2. Are particular items stolen more from particular LGAs and are particular items stolen more times?
  3. Does the type of location area (residential, commercial, and other) have an impact on crime?

Which a clear understanding of the questions I wanted to answer, I moved onto Power BI to start creating charts for the dashboard.

When creating a dashboard in Power BI, I had to consider how I could create a good dashboard for great story telling despite limited interactivity between the three tables that I was not able to join together. I initially decided to create three different views of the tables to resolve the join issue but upon exploring the data more, I realised that it would be difficult to find a story at a high level aggregate detail. Ultimately, I decided to rescope and hone in on looking at property-related crimes in farmland areas of Victoria. I also decided to include another table related to what kind of items were stolen, how many items were stolen and the value of the items. Within this dataset, I noticed a few interesting things such as solar panels, llamas and camels being some of the property that were stolen from the victims. I believed that this would be an interesting story to tell so I created a second view of the dashboard using this table.

Power BI is different from Tableau in that charts for dashboard can be create easily and quickly. However, Tableau offers a higher ceiling for customisation and dashboard interactivity. As a result, for this challenge, I decided that the best course of action would be to focus on what limited functionality I could add to Power BI as well as dashboard design. Due to limitations on publishing the power BI dashboard, I included a screenshot of the final dashboard below. My dashboard also contains two pages which can be reached by clicking on the bookmark tab.

The Data School
Author: The Data School