For Tuesday’s task, Andy asked us to download and analyse daily summary weather observations from the National Weather Service’s Cooperative Observer Program. His specific instructions were:
“The team needs to download weather observations from every station for every day.”
No problem, find the API, work out how to use it to download the lot and BAM.
Slight problem, APIs have limits and this one’s limit is 5 requests per second or 10,000 per day.
Big problem, there’s about 800k weather stations, some of which have daily data for 200+ years.
Even bigger problem, daily summary data can only be downloaded one year at a time, meaning for every weather station, for every year, we would have to make a call to the API.
Adding those three problems together meant we’d be making a call 800k stations, for possibly 200+ years, which would far exceed the number of calls we were allowed. On top of that, Nick started talking about things like iterative macros inside of batch macros, so I decided to call time on the idea of using the API. Luckily, this happened before lunch, so I chose a reasonable number of stations with an idea of a story in mind, and began to download manually.
I ran my data through Alteryx to union and clean up a couple of bits, then opened Tableau. Shortly after this I realise I hadn’t downloaded all the dates I’d meant to for one of the locations, so re-downloaded the data and ran it through again. Then I realised I had missed out some locations, then the wrong something else, so all in all I downloaded my data about five times. Each time, I only realised once I open it in Tableau, so had to return to the workflow, opening and closing my half started workbook each time.
So when Andy announced at half one that we had to use Tableau v.9, everyone groaned at the thought of re-doing what they’d accomplished already. For me though, it wasn’t really that much of a problem for me because I had barely managed to get my data sorted and certainly hadn’t made anything worthwhile anyway.
Finally, around half two, I was done and opened version 9 with all the enthusiasm I could muster. Cue an afternoon of “Oh wow, look at the funny UI” and “So annoying, you can’t use spatial files/ add padding in a dashboard/Viz in Tooltip/where’s Tableau book?”. I might not win any prizes for insights from my finished dashboard, but I’ve certainly had lots of insights of my own into the world of Tableau pre-2018.