I couldn’t tell you why, but one day I decided to check out Google Trends, knowing about it of course, but never having actually used it.
I navigated to the trends page and was met with a search bar. First things first, I entered my name. Zilch. Haha, naturally, I’m not that interesting. Feeling a rumble in my tummy I next entered “Pie”. Quite delighted, I was met with an obvious seasonal pattern.
I ran my cursor over the dates and although I am Australian I could make out that some of the peak searches aligned with American Holidays – Thanksgiving, Independence Day etc. But what was the peak that appeared in March each year? A little googling, and I suspected that it aligned with “Pi Day”. Pi Day is celebrated annually on the 14th of March (in American date format, it would be 3.14 – a rounded π). A celebration I had not heard of before, and whilst this is considered an international celebration, I’m not inclined to be celebrating it anytime soon (I’m far too lazy to be baking pies). Irregardless… this could make an interesting viz.
I looked at the embed options of the trends chart and found that it utilised javascript – might prove a little challenging for a recreational viz. I then peeked at the code and considered copying as SVG and illustrating over the top. Again, a little too involved for a recreational viz. So downloaded the CSV and looked to recreate the chart. I thought about the pulse chart in AppSource, this would be a great way of showing trend and illustrating key points in time. I went however, to Datawrapper.
I uploaded the CSV data to Datawrapper and began creating a line chart. I annotated key dates and highlighted COVID lockdown dates, where the seasonal searches appeared to come undone. Pie searches appear to be higher at this point, perhaps due to more people cooking at home. Happy with the chart, I published and copied the embed code.
Over in Power BI desktop, I pasted the embed code into a measure, dragged the measure into the HTML Content visual, published the report to the service and viola!
Not a data story as yet… but one day I may get around to polishing it off 😊
