Dashboard week started with focus on analysis today. Andy provided us with a great challenge of downloading UN data and letting our inner analyst run wild.
A little late on the shotgun I found myself with North America as my continent. This should be easy right? First world countries with gigabytes of unbiased, historical data from established and well run water quality stations. Nope, not today.
Instead it became a hide and seek adventure, trying to find a variable that may:
- Expose interesting insights
- Have enough historical data to validate an unbiased trend
- And be measured by a considerable amount of stations within the US/Canada.
There were a few iterations starting in Alteryx and looping back after no significant trends found in Tableau. Many investigations later (including Mercury, Total dissolved solids, Lead and even Nitrogen readings), still no significant trend. As well as limited stations reporting these measures.
So using a broader factor like pH I got the result. A trend with a p-value under 0.05. Generalizing with pH seemed to show influences on the quality of surface water over time.
The data prep in Alteryx was simply cleaning the data and creating some spatial points for mapping later in Tableau. Multiple CSVs in, one nice Tableau hyper out.
The final key step was getting the required filters into the dashboard for functionality. As time was limited, practical filtering without flashy tricks was prefered and added the range to show how pH is slowing transforming over time.