Comprehensive Air Quality and Meteorological Dataset with Advanced Statistical and Fractal Features

Description This repository contains a comprehensive, high-resolution time-series dataset (55,272 hourly observations) combining air quality parameters and meteorological data. The dataset is presented in two variants: a raw/collected base dataset and an extended dataset enriched with advanced statistical features, dimensionality reduction components, and non-linear dynamic indicators. This dataset is highly suitable for machine learning forecasting, climate research, environmental pollution modeling, and complex time-series analysis. Dataset Structure & Variants Base Dataset (collected_data.csv): Contains 19 core features including standard air pollutants (AQI, PM2.5, PM10, CO, SO2, NO2, O3) and basic weather conditions (temperature, humidity, precipitation, pressure, wind speed/direction, and weather codes). Extended Dataset (extended_data.csv): Contains 36 features. It includes all base parameters plus newly engineered features: Principal Component Analysis (PCA) indices, Moving Averages (MA), and advanced fractal analysis metrics (Hurst Exponent, Detrended Fluctuation Analysis, and Higuchi Fractal Dimension). Data Dictionary (Column Definitions) Core Air Quality & Meteorological Features (Both Datasets) date: Timestamp of the observation (hourly resolution, datetime64[ns]). AQI: Air Quality Index (integer). pm2_5 / pm10: Particulate matter concentrations (less than 2.5 micrometers and less than 10 micrometers, respectively). CO / SO2 / NO2 / O3: Concentrations of Carbon Monoxide, Sulfur Dioxide, Nitrogen Dioxide, and Ozone. temperature_2m: Air temperature at 2 meters above the ground (Celsius). relative_humidity_2m: Relative humidity at 2 meters (percentage). apparent_temperature: Perceived temperature combining wind chill and heat index. precipitation / rain / snowfall: Total precipitation, liquid rain, and snow water equivalent measurements. surface_pressure: Atmospheric pressure at the surface. cloud_cover: Total cloud cover percentage. wind_speed_10m / wind_direction_10m: Wind speed and wind direction at 10 meters above the ground. weather_code: Categorical code representing current weather conditions (e.g., WMO weather codes). Engineered & Statistical Features (Extended Dataset Only) pollution_index_pca: First principal component extracted from air pollutant features representing general pollution intensity. weather_index_pca: Principal component extracted from meteorological features representing generalized weather states. [Feature]_ma: 24-hour Moving Average (smooths short-term fluctuations for AQI, pollution PCA, and weather PCA). [Feature]_hurst_exponent: Hurst exponent (H) calculated over rolling windows, indicating long-term memory/persistence of the time series. [Feature]_dfa: Detrended Fluctuation Analysis scaling exponent, indicating long-range power-law correlations. [Feature]_hfd: Higuchi Fractal Dimension, representing the complexity and roughness of the time-series signal. [Feature]_fractal_dimension: Estimated fractal dimension of the signal trajectory. Note on Missing Values: The advanced statistical and rolling indicators (columns 21–35) contain a small fraction of missing values at the beginning of the time series due to the required calculation window sizes (rolling lookback constraints).