CO₂ Time Series Analysis

Time series analysis and forecasting of atmospheric CO₂ concentration measured at Mauna Loa Observatory.

---

Project Overview

This project performs a complete time series analysis on atmospheric CO₂ concentration.
The goal is to explore long-term trends, seasonality patterns, stationarity, and to build an ARIMA-based forecasting model.

The workflow includes:

• Loading and preprocessing the CO₂ dataset
• Monthly resampling
• Rolling mean and rolling standard deviation analysis
• Seasonal decomposition (trend, seasonal, residual)
• ADF stationarity test
• ACF & PACF analysis
• SARIMA modeling
• Forecasting and visualization

This project is implemented in Python and can be run locally.

---

Files in This Repository

• CO2_Analysis.py — Main analysis script (time series pipeline)
• co2.csv — Raw CO₂ dataset (daily measurements)
• README.md — Project documentation

---

Dataset Description & Source

The dataset used in this project contains daily atmospheric CO₂ measurements collected at:

Mauna Loa Observatory (MLO), Hawaii
Operated by the NOAA Global Monitoring Laboratory (NOAA-GML).

Official Data Source:

NOAA Earth System Research Laboratories (ESRL)
Global Monitoring Laboratory – CO₂ Trends

Website:
https://gml.noaa.gov/ccgg/trends/

The dataset in this repository (co2.csv) is derived from the NOAA open dataset and formatted for local time series analysis.
NOAA provides this data publicly for research and educational use.

---

How to Run Locally

1. Install dependencies

pip install pandas numpy matplotlib statsmodels scipy

2. Run the script

python CO2_Analysis.py

The script will automatically:

• Load co2.csv
• Convert the date format
• Resample to monthly means
• Generate visualizations
• Fit a SARIMA model
• Produce CO₂ forecast plots

---

Output Visualizations

Running the script produces:

• Monthly CO₂ concentration plot
• Rolling mean / rolling standard deviation
• Seasonal decomposition (trend, seasonal, residual)
• ACF graph
• PACF graph
• SARIMA model diagnostic plots
• Forecast vs actual plot with confidence intervals

These figures reveal long-term atmospheric CO₂ trends and provide predictive insight.

---

License

This project uses publicly available data from NOAA-GML.
Please cite NOAA appropriately if you use this data in other analyses or publications.

---

Acknowledgments

Special thanks to:

• NOAA Global Monitoring Laboratory for providing long-term atmospheric CO₂ records
• Researchers at Mauna Loa Observatory for continuous CO₂ monitoring since 1958