air-quality-tests
Working with air quality datasets from many sources. Tools: Excel, .NET 8, Visual Studio
About the dataset
Sources
Official
- Buenos Aires Data - Calidad de Aire
- Calidad del Aire | Buenos Aires Ciudad - Gobierno de la Ciudad Autónoma de Buenos Aires
- Estación Parque Centenario | Buenos Aires Ciudad - Gobierno de la Ciudad Autónoma de Buenos Aires
- Red de Monitoreo - Buenos Aires Ciudad - Calidad Ambiental
- Obtener contaminantes por día por lugar - Buenos Aires Ciudad - Calidad Ambiental
External providers
Noticias
- Incendios y calidad del aire: Un vínculo inquietante en la Ciudad de Buenos Aires | Sobre La Tierra (uba.ar)
- Especialistas de la UBA alertan por la calidad del aire en la Ciudad (ambito.com)
Data characteristics
Ficha de contaminante - CO.
Ficha de contaminante - NO2.
Ficha de contaminante - PM10.
Fuentes recolectoras de datos.
Transforming data
- Calculating AQI (Air Quality Index) Tutorial (kaggle.com)
- Add missing fields: season y "color" foreach contaminant (based on previously defined effects).
Analysis approach
| Combination | Approach | Use Case | Benefits |
|---|---|---|---|
| Clustering + Decision Trees | Cluster by pollutant levels across locations, apply decision trees to predict air quality. | Group air quality patterns by time/location and predict poor air quality. | Reveals patterns and creates predictive rules for specific clusters. |
| Decision Trees + Association Rules | Use decision trees to identify key pollutants, apply association rules to find pollutant co-occurrences. | Identify key pollutants driving poor air quality, and find co-occurrence patterns. | Uncovers important pollutants and frequent co-occurrences. |
| Clustering + Association Rules | Cluster based on pollutant levels at different locations, then apply association rules within each cluster. | Understand pollutant combinations that spike together across locations. | Targeted pattern discovery for specific clusters of air quality data. |
| Clustering + Decision Trees + Association Rules | Cluster data, use decision trees to predict air quality, and mine association rules within clusters. | Comprehensive air quality analysis across time and locations. | Multi-faceted approach combining segmentation, prediction, and patterns. |
| Association Rules Informed by Decision Trees | Use decision trees to find key pollutants, then apply association rules on these pollutants. | Focus on critical pollutants and discover related pollutant patterns. | Reduces complexity by focusing on the most important features. |
Architecture Records
1. On DB Injection
Context
Based on this file for a current sample: aspire-samples/samples/AspireShop/AspireShop.BasketService/AspireShop.BasketService.csproj at main · dotnet/aspire-samples (github.com)
Decision
I will inject the DB from the AirQualityApp.ApiService project
2. Processing Application
Context
Based on experience, the main objective of the app will be exposing through an API/web frontend, the consumed/processed dataset
Decision
The architecture will be like:
3. CSV Reading approach
Context
My options are:
- Manual parsing.
JoshClose/CsvHelperNuGet Package.
Decision
We are using this NuGet package: JoshClose/CsvHelper: Library to help reading and writing CSV files (github.com)
CsvHelper Getting started