Not that I should be avoiding steps, but I was tired of getting up to walk over to my Ambient Weather WS-5000 (or 4000 or 2000) console (or it's Ecowitt HP2561 sibling) to view the current weather outside. So with the help of Gemini and CODEX I put together a dashboard (app, device, and dashboard) that is visually similar, offering all of the same data (and just a little more).
My setup was originally an Ambient Weather WS-4000 with both indoor and outdoor air quality sensors, lightning detection, monitoring for all of our refrigerators and freezers along with a pool water temp sensor. I added to that a Ecowitt HP2561 to capture all of the same data and push that into Hubitat via the Ecowitt WiFi Gateway driver.
My Weather Dashboard App collects that ecowitt data, tracks trends, builds a near term forecast and then pushes formatted data to a dedicated Weather Dashboard Driver that is used by a Javascript custom dashboard tile. The Javascript dasbboard tile can be loaded using a Hubitat Dashboard or directly in any local network browser by loading the html file you will find in your File Manager (requires Maker API).
Despite nearly 30 years across commercial product engineering and InfoSec, I used AI to build the project; Gemini at first, but Codex ultimately to rebuild and polish. The goal was to force myself to be only the architect, driving AI to build a high quality product. Plus I wanted my weather data in a network accessible pretty format.
You can install it using HPM, Install -> URL -> https://raw.githubusercontent.com/sonoranwanderer/hubitat_weather_dashboard/main/packageManifest.json
Full setup and configuration instructions are at the GitHub repository Hubitat Weather Dashboard
This dashboard can actually use any weather data source so long as the data is exposed by attributes using 1 or more devices on your Hubitat system. Data mapping instructions are at the link above, the default mapping assumes Ecowitt WiFi Gateway. The dashboard appearance is customizable, order and sizes, including hiding sections you do not have the equipment for (eg Air Quality, lightning).
You can also interact with it, note the inline buttons for conversion of standard to metric units, and pause/resume and quick cycling of the ambient sensors.
Hate it? Love it? Think it could be a lot better in some particular way? Let me know.
Footnote: the forecast feature will report immediately, but requires some time to build the dataset to become useful. The app does the forecasting and keeps a rolling 30 day internal history. It's a simple, common algorithm.
