Introducing Power BI

Day 3 brings us a very different challenge as we are tasked with using Power BI to build our dashboards. Today’s data comes from the City of New Orleans open database, covering 311 call history (the local government’s information and non-emergency call number).

The first hurdle was to access the data through an API, which presented an initial challenge due to ambiguous documentation. But after adopting a trial-and-error approach, I was able to download just over 300,000 rows of call data.

‘Chuck it in’

That’s probably the best way to describe my Power BI dashboarding approach today. After spending an hour or two exploring the interface and functionality of the software, I decided that probably its key strength is in enabling people to conduct analysis very quickly. The medium to create viz is through a slightly more developed version of Tableau’s ‘Show Me’. Throw the data in, select a chart type, select a few more filters – and voila. The charts are far less customisable, but if that’s not really a concern and you just need to put some charts together on the fly, then Power BI does the job very well.

I was also impressed by the ‘Analyze’ feature, which actually suggested a relevant and useful chart in the context of my work (examining average completion times for 311 calls). It was in exploring the software’s suggestions that I thought of ideas for charts that would help a user identify patterns and trends in the data.

The ‘Analyze’ data feature

I chose to keep the dashboard exploratory, meaning it’s also quite busy with 5 charts, but Power BI made it simple for me to do so: there was no having to mess around with containers to get things to align and no worrying about dashboard actions to enable interactivity. In the space of a couple of hours, I was able to create a functional dashboard allowing me to interrogate most aspects of the dataset – and it doesn’t look too bad either. I can imagine it would be much harder to create something that would rival Tableau in the beauty stakes, and any advanced/custom charts that aren’t in the gallery are out of reach. But it’s a tool that gets you quite far, quite fast with minimal faf-time. Of course, it’s some of that faf that makes Tableau a more powerful tool for customisation.

The final outcome