Dashboard Week - Day 2
This is my final week in training, and my cohort and I are tasked with creating a dashboard every day. Today's task was to create a dashboard using the Numbers API, which generates facts from different categories. It sounded simple enough, but quickly became three different iterations before I found a proper focus.
The first challenge was making the API call, which wasn’t too tricky thanks to the documentation. The real challenge was deciding how to shape the data into something meaningful. My first idea was to generate a row per fact type and make an API call that produced random facts. But given the scope, I soon realised this approach wasn’t the best use of time.
Next, I tried generating the first 100 rows by doing a batch call and specifying a range of data. Once I had that, I cleaned and formatted the data in Power Query, splitting it into rows and then columns and repeating those steps for each table. I also had to do some extra transformations and calculated columns, which I didn’t anticipate needing until I got into building visuals.
Initially, I planned to make a few bar charts to see if anything interesting stood out, but it felt unengaging. Instead, I shifted towards an interactive dashboard where users could select a fact type, and a custom measure would generate a random fact. I added another feature to explore facts by month, which meant another round of custom columns in Power Query. By the end of the day, I had two dashboards, one for interactive exploration and one listing all facts by type.
The biggest challenge was creating a refresh button that users could click to get new facts. I experimented with a slicer and parameter to cycle through facts, but it wasn’t intuitive. I even renamed the slicer to clarify, but it didn’t feel right. Technically, it’s possible to implement a proper refresh, but it was too time-consuming to fit into the day, so I stuck with the slicer.
One unexpected twist came when I actually looked through the data. Many of the facts were about violence, crime, and conflict. But this makes sense, considering many facts are drawn from events over the past hundred years.
Overall, the API call itself wasn’t an issue, but the vagueness of the brief made it harder to land on a focus. I bounced between three dashboard ideas, spent too long wrestling with the refresh idea, and had less time to polish the formatting. That said, I’m still happy with what I produced. Moving forward, I know I need to be quicker in choosing a direction and let go of ideas that take too much time to implement.