Tue
3 Nov
21° / 8.9°
42.5–87.4%
Daily temperature range and humidity for Launceston · Local seasonal model fallback.
The weather provider was unavailable, so this view uses the local seasonal model. Icons summarise its temperature and humidity profile.
The interface separates practical live forecasting from the project’s reproducible IoT and machine-learning baseline. This keeps everyday weather guidance useful without hiding how the original model and data pipeline work.
Daily temperature, humidity, and weather-code guidance come from Open-Meteo’s location-matched forecast feed. The app requests 11 days once, uses today for the dashboard summary, and displays the next 10 days.
Four PHP-ML least-squares regressions estimate daily minimum/maximum temperature and humidity. Features include seasonal sine/cosine terms, site indicators, and site-by-season interactions for five Tasmanian locations.
The raw sensor file has over one million half-hourly readings. A compact seasonal cache is generated when the data changes, so visitors do not wait for the full CSV to be scanned on every request.
The accuracy table compares the local seasonal baseline with the same-day provider reference; it is a transparent indicator, not a formal held-out validation score. The original 2015–2021 sensor data is supplemented by a synthetic 2022–2026 extension, so results should be treated as project evidence rather than operational weather guarantees. View the source and methodology on GitHub .