🟢 Time Series Forecasting with ARIMA – Predicting Daily Minimum Temperatures

AIVIA Project Engine Recommender & Decision Systems
, , , , , ,
1

Time Series Forecasting with ARIMA – Predicting Daily Minimum Temperatures

Objective

Build a simple time series forecasting model to predict daily minimum temperatures. Learn fundamental time series concepts and basic statistical forecasting techniques.


Learning Outcomes

By completing this project, you will:

  • Understand time series analysis basics
  • Implement simple statistical forecasting models
  • Use essential time series libraries and tools
  • Visualize and interpret time series data
  • Evaluate model performance with basic metrics

Prerequisites

  • Basic Python programming skills
  • Basic statistics knowledge
  • Familiarity with pandas and NumPy
  • Basic understanding of plotting with matplotlib

Dataset Details

  • Daily Minimum Temperatures Dataset: Link
  • Size: ~20KB
  • 10 years of daily temperatures from Melbourne (1981–1990)
  • Features: Date, Temperature (°C)

Tools Required

# Core libraries
pip install pandas numpy matplotlib

# Statistical modeling
pip install statsmodels

Project Structure

temperature_forecasting/
│
├── data/
│   └── daily-min-temperatures.csv
│
├── src/
│   ├── data_processing.py
│   ├── model.py
│   └── evaluation.py
│
└── notebooks/
    ├── data_exploration.ipynb
    ├── model_implementation.ipynb
    └── model_evaluation.ipynb

Steps and Tasks

1. Data Acquisition and Exploration

Load and explore the dataset.

import pandas as pd
import matplotlib.pyplot as plt

# Load data
df = pd.read_csv('daily-min-temperatures.csv', parse_dates=['Date'], index_col='Date')

# Plot time series
plt.figure(figsize=(12, 6))
plt.plot(df['Temp'])
plt.title('Daily Minimum Temperatures in Melbourne')
plt.xlabel('Date')
plt.ylabel('Temperature (°C)')
plt.show()

2. Basic Statistical Model Implementation

Implement a simple forecasting model using Moving Average.

# Calculate moving average
df['Moving_Average'] = df['Temp'].rolling(window=365).mean()

# Plot moving average
plt.figure(figsize=(12, 6))
plt.plot(df['Temp'], label='Original')
plt.plot(df['Moving_Average'], label='365-Day Moving Average', color='red')
plt.legend()
plt.show()

3. Forecasting with ARIMA

Use ARIMA for forecasting.

from statsmodels.tsa.arima.model import ARIMA

# Split data
train = df['Temp'][:'1989']
test = df['Temp']['1990':]

# Fit ARIMA model
model = ARIMA(train, order=(5,1,0))
model_fit = model.fit()

# Forecast
forecast = model_fit.forecast(steps=len(test))

# Plot forecast vs actual
plt.figure(figsize=(12, 6))
plt.plot(test.index, test, label='Actual')
plt.plot(test.index, forecast, label='Forecast', color='red')
plt.legend()
plt.show()

4. Model Evaluation

Evaluate the model’s performance.

from sklearn.metrics import mean_absolute_error

mae = mean_absolute_error(test, forecast)
print(f'Mean Absolute Error: {mae:.2f}°C')