Is it Windy Outside?

Or, What's Up With The Great Lakes Image?

NOTE: The blog header image has been replaced by Clifford attractors. A static archived wind map can be seen here

My earlier experience with Wordpress made me think there had to be a better way to do simple, reliable sites online. A blog is static, meaning everyone sees the same content. Yihui Xie from RStudio created what for my needs is a better way to host content - blogdown.

The web is built with HTML, Javascript and css. Everything boils down to these eventually. But that doesn’t mean you have to use these languages directly to make sites for the web. R Markdown is a much simpler markup language that makes it easy to format text and basic tables. Add in Hugo, a very fast framework for combining markdown with pre-existing themes, and a simple static web site that looks good, requires fewer resources and is faster than Wordpress is within reach.

Yihui cautioned about the lure of fancy themes and how boredom will settle in, no matter what, in time. I agree. Nevertheless I chose, the beautifulhugo theme which I classify as minimally fancy. I added a semi-dynamic header image that updates every three hours with current wind conditions over the great lakes region. Finally, an answer is at hand to humanity’s age-old question, “Is it windy outside?”

The wind data is from National Oceanic and Atmospheric Administration (NOAA). The beautiful watercolour maps are from Stamen Design in the Bay Area.

I also followed Yihui’s advice to use Github and Netlify to host this blog. Who can argue with a free, global content delivery network (CDN). CDNs are basically racks of computers distributed around the globe to store content closer to the end-user. And it is fast!

But wait you say, how do you get changing content on a static site? I needed a reliable server to fetch the wind data and build a new image every three hours. Unfortunately, my current host for lpphoto.me doesn’t allow shell access without a static IP address, which is cost prohibitive for what I require.

I decided to check out Google’s Compute Engine and soon had R running on a small virtual machine in the cloud. After the image is created it’s moved to a bucket in Google’s Cloud Storage which is where the blog finds it.

Google’s service is more than I require and I will probably migrate to Digital Ocean in the future. This explains why.