Isaac-Ndirangu-Muturi-749/Time-Series-Forecasting-for-Electric-Power-Consumption

This project employs time series forecasting on household electric power consumption data spanning 47 months. Using univariate modeling with Facebook Prophet, the analysis addresses missing values and extends to multivariate forecasting. Evaluation metrics such as RMSE and MAE gauge model performance.

Time Series Forecasting for Electric Power Consumption

Project Overview

This project focuses on time series forecasting for individual household electric power consumption. The dataset spans December 2006 to November 2010, providing a rich source of information for understanding and predicting energy consumption patterns.

Project Structure

The project is organized into several key components:

1. Notebook Sections:

   • exploratory_data_analysis: Explore the dataset, handle missing values, and visualize time series data.
   • time_series_decomposition: Decompose the time series into trend, seasonality, and residual components.
   • arima_modeling: Apply ARIMA modeling for time series forecasting.
   • prophet_univariate_model: Utilize Facebook Prophet for univariate time series modeling.
   • prophet_multivariate_model: Extend modeling to multivariate forecasting using Facebook Prophet.
   • evaluation_metrics: Evaluate model performance using metrics such as RMSE, MAE, and MAPE.

2. Dependencies:

   • The project dependencies are listed in the requirements.txt file.
   • Install dependencies using: pip install -r requirements.txt

3. Data:

   • The dataset is sourced from the UCI Machine Learning Repository.
   • Preprocessing includes handling missing values and converting columns to numeric.

Attribute Information

• date: Date in dd/mm/yyyy format
• time: Time in hh:mm:ss format
• globalactivepower: Household global minute-averaged active power (in kilowatt)
• globalreactivepower: Household global minute-averaged reactive power (in kilowatt)
• voltage: Minute-averaged voltage (in volt)
• global_intensity: Household global minute-averaged current intensity (in ampere)
• submetering1: Energy sub-metering No. 1 (watt-hour of active energy)
• submetering2: Energy sub-metering No. 2 (watt-hour of active energy)
• submetering3: Energy sub-metering No. 3 (watt-hour of active energy)

4. Modeling:

   • ARIMA Modeling: Time series modeling using ARIMA (AutoRegressive Integrated Moving Average) for forecasting.
   • Facebook Prophet: Utilizing the Prophet library for univariate and multivariate time series forecasting.

5. Evaluation:

   • Performance metrics include Root Mean Squared Error (RMSE), Mean Absolute Error (MAE), and Mean Absolute Percentage Error (MAPE).

Project Outcomes

• Insights into Electric Power Consumption:

  • Understand patterns, trends, and seasonality in household power consumption.
  • Identify factors contributing to energy usage.

• Time Series Modeling:

  • Apply ARIMA modeling to capture time-dependent patterns.
  • Utilize Facebook Prophet for both univariate and multivariate forecasting.

• Performance Evaluation:

  • Assess model accuracy using RMSE, MAE, and MAPE.
  • Understand the effectiveness of different forecasting approaches.

• GitHub Repository:

  • Access the full project code and documentation on GitHub.

Contact Information

Follow me on Twitter 🐦, connect with me on LinkedIn 🔗, and check out my GitHub 🐙. You won't be disappointed!

👉 Twitter: https://twitter.com/NdiranguMuturi1
👉 LinkedIn: https://www.linkedin.com/in/isaac-muturi-3b6b2b237
👉 GitHub: https://github.com/Isaac-Ndirangu-Muturi-749

Feel free to reach out for questions or collaborations!