AQI Don't Wanna Die is powered by a custom machine-learning model using binary Support Vector Classification.
After the appropriate model selection and h-param tuning, our model achieved an accuracy of 96% on a test set.
(The test set represented the full population of our data, using stratify-shuffle-split for an accurate split of "safe" vs "unsafe" classifications).
AQI, or Air Quality Index, is a numerical scale that runs from 0 to 500 and describes the level of air pollution in a specific area. Our alternative, AQI Don't Wanna Die, provides a more straightfoward approach with a "safe" or "unsafe" label, and lets users see and modify crucial factors.
We live in a world that provides many resources, from the water we drink, the food we eat, and the air we breathe. Unfortunately, some of these assets may become tainted over time. Most recently, Seattleites experienced particularly poor conditions, having the worst air quality in the world on consecutive days. In response to these events, we created this project to keep people safe from hazardous surroundings and spread awareness about air pollution.
Front-end web development
ML engineering + deployment
Front-end web development