Weather Dashboard: ten AccuWeather features in Tkinter

Published:

A desktop weather dashboard built on the AccuWeather API, written as teaching material for KAGS381 (Python). It grew out of the course’s accu_*.py exercises: the same API calls, arranged into something students can read one file at a time.

View the project on GitHub

What it shows

Three tabs — Now, 12 hours and Compare — covering ten features:

  • Heat stress — wet-bulb globe temperature on a coloured gauge, with “feels like” in sun and shade
  • Umbrella? — a YES / MAYBE / NO verdict from the next 12 hours of rain and thunderstorm chance
  • Weather theme — icons and card colours chosen by AccuWeather’s icon number, so they survive a language change
  • Headline — the most significant weather in the next five days, coloured by severity
  • Sun & moon — the sun’s position now along its arc, plus the moon’s phase drawn from its age
  • Wind & pressure — a compass arrow pointing where the wind blows to, with the pressure trend
  • vs yesterday — warmer or cooler than 24 hours ago, without needing the paid history endpoint
  • Best hour — each of the next 12 hours scored for going outside, with the best window highlighted
  • Animated sky — rain, wind-slanted streaks, lightning, clouds and day/night, hour by hour
  • Compare cities — up to four cities side by side on one shared temperature scale

How it is built for teaching

  • One networking module. accu_api.py is the only file that touches the network; everything else receives a plain dictionary. Students can read any panel without thinking about HTTP.
  • Every panel runs on its own. python3 panel_wind.py opens just the compass, with demo data.
  • It works with no API key. Without one the app plays built-in sample weather, marked with a DEMO badge, so the code can be read and run before signing up for anything.
  • 250 tests, none touching the network. The API is replaced with fakes, so the suite runs offline in about two seconds.
  • The key never goes in a file. It is read from an environment variable and sent over HTTPS in an Authorization header — the habit the earlier course scripts were written to break.

The image above is drawn by the app itself: the storm hour of the animated scene, the heat gauge, the wind compass, the sun arc, and the twelve scored hours of a Bangkok afternoon.