Predicting Air Quality Index using Python

Air pollution is a growing concern globally, and with increasing industrialization and urbanization, it becomes crucial to monitor and predict air quality in real-time. One of the most reliable ways to quantify air pollution is by calculating the Air Quality Index (AQI). In this article, we will explore how to predict AQI using Python, leveraging data science tools and machine learning algorithms.

What is AQI?

The Air Quality Index (AQI) is a standardized indicator used to communicate how polluted the air currently is or how polluted it is forecast to become. The AQI is calculated based on pollutants such as:

• PM2.5
• PM10
• NO2
• SO2
• CO
• O3

Each pollutant has a sub-index, and the highest sub-index among them becomes the AQI.

I = \frac{I_{HI} - I_{LO}}{BP_{HI} - BP_{LO}} \times (C - BP_{LO}) + I_{LO}

Where:

• I is the AQI
• C is the concentration of the pollutant
• BP_{HI}, BP_{LO} are the breakpoint concentrations
• I_{HI}, I_{LO} are the AQI values corresponding to those breakpoints

We can see how air pollution is by looking at the AQI

Let's find the AQI based on Chemical pollutants using Machine Learning Concept. 

Data Set Description

It contains 7 attributes, of which 6 are chemical pollution quantities and one is Air Quality Index. AQI Value, CO AQI Value, Ozone AQI Value, NO2 AQI Value, PM2.5 AQI Value, lat,LNG are independent attributes. air_quality_index is a dependent attribute. Since air_quality_index is calculated based on the 7 attributes.

As the data is numeric and there are no missing values in the data, so no preprocessing is required. Our goal is to predict the AQI, so this task is either Classification or regression. So as our class label is continuous, regression technique is required.

Step-by-Step Process to Predict AQI

1. Importing Libraries

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import seaborn as sns
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.ensemble import RandomForestRegressor
from sklearn.metrics import mean_absolute_error, mean_squared_error, r2_score

2. Loading the Dataset

We’ll use a dataset with pollutant concentration levels and corresponding AQI values.

data = pd.read_csv('air_quality_data.csv')
print(data.head())

3. Data Preprocessing

Handle missing values, rename columns, and check data types.

data = data.dropna()
data.columns = [col.strip().lower() for col in data.columns]

4. Exploratory Data Analysis (EDA)

Visualizing relationships between variables.

sns.pairplot(data)
plt.show()

corr = data.corr()
sns.heatmap(corr, annot=True, cmap='coolwarm')

5. Feature Selection

Choose relevant features for training.

X = data[['co aqi value', 'ozone aqi value', 'no2 aqi value', 'pm2.5 aqi value']]
y = data['aqi value']

6. Train-Test Split

X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)

7. Model Training (Random Forest)

model = RandomForestRegressor(n_estimators=100, random_state=42)
model.fit(X_train, y_train)

8. Model Evaluation

y_pred = model.predict(X_test)

print("Mean Absolute Error:", mean_absolute_error(y_test, y_pred))
print("Mean Squared Error:", mean_squared_error(y_test, y_pred))
print("R2 Score:", r2_score(y_test, y_pred))

9. Plotting Results

plt.figure(figsize=(10, 6))
plt.plot(y_test.values, label='Actual AQI')
plt.plot(y_pred, label='Predicted AQI', alpha=0.7)
plt.title('Actual vs Predicted AQI')
plt.legend()
plt.show()

Output:

Model Evaluation Metrics: Mean Absolute Error: 0.09 Mean Squared Error: 2.59 R2 Score: 1.00

Real-world Applications

• Smart cities to monitor pollution in real-time.
• Healthcare apps to warn sensitive populations.
• Environmental agencies for policy formulation.

Dataset Link: click here.