Dashboard Week Day 3 – New Orleans 311 Calls & Power BI fun

by Brian Scally

As is becoming tradition, DS12 had to use Microsoft Power BI for one of the Dashboard Week projects. Today was the day, and we were given the New Orleans 311 call database to work with.

The database contains all kinds of public requests and complaints about non-emergencies, such as complaints about potholes, graffiti, dead animals on the roads, illegal dumping, etc. Immediately I decided to reduce the data to one topic. Initially I started looking at the requests for the removal of dead animals, but after working with Power BI for 30 minutes I had reached such a level of unhappiness and discomfort that I had to go back and change topic. In the end I went with rodent complaints, which I considered to be far less grim.

Along the way I tried to document some features of Power BI that I came across that I either liked or disliked. So in my completely biased opinion, having gotten to learn Tableau inside out for the last three months, here they are.


  • Cool interactivity and automatic filters.
  • Automatic proportional highlighting.
  • Good looking animations when switching between filters, feels professional.
  • Of the few bits of Power Query that I used, I found it to be relatively intuitive.
  • Ability to add custom visuals from the Microsoft Store, though to be honest by the looks of them, I would never use half of them.
  • VIOLIN PLOTS (via the custom visuals – COME ON Tableau).
  • ArcGIS maps – didn’t get to play much but they look quite good.


  • No native histogram tool, or intuitive functions to make bins without calculations (as far as I’m aware). The histogram tool I downloaded was fairly awful. You have to define the number of bins rather than the bin size, which is really unintuitive.
  • You can’t edit titles and axes by double clicking on them, you have to go into a menu.
  • I found myself squinting a lot at the Vizualisations pane, trying to figure out where to drop what.
  • To make a BAN, you have to use something called the ‘Card’ visualisation, which has a title that you can’t fully customise.
  • Can’t customise tooltips
  • The native filled maps look poor if you colour by a measure, and I couldn’t find out how to make the colours opaque.
  • I also had a bad time with the dotted maps – the sizing of the dots is really inflexible, I couldn’t get them small enough to not majorly overlap.
  • You also can’t wash out the background map, if you wanted to just show filled regions like zip codes.

“But Brian, it let’s you display plots that you build with R code! You love R code!”

Sure, you can make some pretty things in R with ggplot2, but when it comes to sticking a ggplot2 visual into a dashboard that likely has completely different formatting, not to mention losing all that cool interactivity between charts on the dashboard (one of the only things I really thought was awesome about Power BI), I think I’ll pass.

I could probably go on, but those were the main points. If you’ve made it this far then you can guess that I’m not exactly devastated to go back to Tableau tomorrow.

Here’s the product of what I managed to put together today, complete with unformattable tooltips, uninterpretable histogram bin sizes, and uncompromisingly transparent filled map, for some visual references to the points I mentioned earlier.

© 2022 The Information Lab Ltd. All rights reserved.