RefinAir

RefinAir — Air quality data, analysis and IoT tools for Dhaka

This repository collects datasets, analysis scripts, Arduino/ESP firmware and diagrams used for PM2.5 and aerosol analysis, monitoring prototypes, and documentation produced during the RefinAir project.

Purpose: provide cleaned datasets, reproducible Python analysis and visualization code, and Arduino/ESP examples for local air-quality logging and uploads.

Repository Structure

• PM2.5 Data/: Time-series PM2.5 CSVs (2017–2021) used for analysis and plotting.
• IQAIR Data/: Raw AirVisual/IQAir data dumps (monthly text files).
• Arduino_Code/: Main Arduino sketch(s). See Arduino_Code/Arduino_Code.ino.
• Arduino Mega Wifi/: Firmware, libraries and flashing tools for ESP8266/ESP modules and Mega Wifi builds.
• Pollution Analysis - python/: Python scripts, notebooks and helper modules used to generate figures and perform modeling.
• Python Datalogger/: Data ingestion and datalogger scripts (serial/HTTP upload helpers).
• Diagrams/: BPMN, maps, schematics and other Visio diagrams used in reports.
• Overleaf/: LaTeX source (paper/thesis) and bibliography.
• Top-level CSV files: important single-file datasets (e.g. climate_historical_data_2017-2021Jun.csv, dhaka-us consulate-air-quality.csv, satellite AOD files g4.areaAvgTimeSeries...csv).
• LICENSE: repository license.

Quick Start (Windows / PowerShell)

• Requirements: Python 3.8+, pip, optionally virtualenv or venv, and the Arduino IDE (or PlatformIO) for microcontroller code.
• Recommended Python packages: pandas, numpy, matplotlib, seaborn, scikit-learn, jupyterlab. Create a requirements.txt in Pollution Analysis - python/ if you want reproducible installs.

Create and activate a virtual environment (PowerShell):

python -m venv .venv; .\.venv\Scripts\Activate.ps1
pip install --upgrade pip
pip install pandas numpy matplotlib seaborn scikit-learn jupyterlab

Start JupyterLab (to open notebooks in Pollution Analysis - python/):

jupyter lab

Running the analyses

• Open notebooks or scripts in Pollution Analysis - python/ and run cells top-to-bottom. Data files are referenced relative to the repository root in the PM2.5 Data/ and top-level CSVs.
• If a script expects cleaned CSVs, run any included preprocessing (look for filenames like clean_*.py or notebook cells named "Preprocessing").

Hardware / Arduino notes

• Open Arduino_Code/Arduino_Code.ino in the Arduino IDE.
• Required libraries are included or referenced in Arduino Mega Wifi/libraries/ — install these into your Arduino libraries folder if the IDE shows missing libraries.
• To flash ESP8266/ESP devices, you can use the provided flasher tools in Arduino Mega Wifi/ESP8266_flasher_* or use the Arduino IDE's built-in upload.

Data descriptions (high level)

• PM2.5 Data/: Ground-based PM2.5 measurements aggregated by year/month.
• dhaka-us consulate*.csv: US consulate monitoring station data (public reference station for Dhaka).
• climate_historical_data_2017-2021Jun.csv: Climate covariates used for correlation analysis.
• g4.areaAvgTimeSeries*.csv: MODIS / aerosol optical depth (AOD) region-average time series (satellite-derived aerosol proxy).

Thesis & Research Paper

• Cloud Driven IoT based Big Data Solution_ Air Quality Monitoring System and Predictive Analysis of Satellite Data using Machine Learning.pdf
  Main thesis/research paper documenting the RefinAir project: system architecture, IoT hardware design, data collection methodology, machine learning models for PM2.5 prediction from satellite AOD and ground sensors, and results/evaluation for air quality monitoring in Dhaka, Bangladesh.

Reproducibility tips

• Keep data and code versions together: if you modify a dataset, save a new CSV with a clear suffix (e.g. *_v2.csv).
• Create a requirements.txt in Pollution Analysis - python/ and pin package versions for reproducible environments.
• Add short README or usage notes inside subfolders that contain scripts (e.g. Pollution Analysis - python/README.md) to document how to run specific notebooks.

Contributing

• Fork, create a branch, make changes and open a pull request. Describe changes and data provenance in the PR.

License & Credits

• See the top-level LICENSE file for license details and reuse terms.

Contact / Author

• Repository maintained by the RefinAir project contributors. For questions or to collaborate, open an issue or contact the original author listed in the project files.