Leaving my apartment used to involve a ritual of checking two apps: Android Weather to decide what to wear, and Citymapper to see when the next trains were arriving. Now though, I just glance at my fridge.

A smart fridge? Nope – far from it. But it’s a fridge with a very smart magnet, a 64x32 RGB LED matrix driven by a Raspberry Pi.

I designed the display from scratch to show information I want to know before heading out: current and forecast weather, along with upcoming train departures from my local subway station. This display eliminates nearly all of the “use cases” where I had to check my phone on the way out the door.

I’m pulling data from the free weather.gov API and New York MTA realtime data feeds with a Node.js app. I then generate the image to show on the matrix using node-canvas, which implements the HTML Canvas API.

This project posed unique design challenges in how to convey information within the constraints of a low-resolution display, like designing tiny weather icons that are still understandable. If there’s rain, for example, a degree of cloud cover is implied so I omit the cloud icon.

Since I wasn’t using a graphing library, I also had to implement details like positioning the high and low temperature forecast labels right above their corresponding peaks and valleys on the line graph. This required roughly 200 lines of code to identify increasing and decreasing temperature intervals, center labels, and check for pixel collisions.

The end product is a colorful conversation starter that helps me get out the door quickly and prepared.