nongzhanguan air quality analysis

The machine learning industry is currently obsessed with predictive hype. I built this analysis to prove that rigorous exploratory data science is often more valuable than a black box model. The goal was to take raw environmental data from Nongzhanguan and surface patterns that actually matter for public health monitoring.

I built a notebook-based workflow to clean and interpret years of pollutant and weather measurements. The real challenge was handling the missing values without destroying the underlying time-series signals. By meticulously documenting the imputation choices, the analysis remains grounded in reality rather than statistical guesswork. The biggest takeaway is that descriptive analytics requires deep domain context. Turning a raw CSV into a readable time-series visualization is easy, but translating those seasonal pollution spikes into actionable public health insights is hard. This project reinforced that data science is useless unless it tells a clear, interpretable story.

Repository

Nongzhanguan Air Quality Analysis