Energy Consumption Analysis

Project Overview

This project presents a comprehensive analysis of residential electricity usage across 90,000 daily records. It explores how household characteristics (like size and AC ownership) and environmental conditions (such as temperature) affect energy consumption patterns.

The primary goal is to derive actionable insights for energy providers, policymakers, and sustainability consultants looking to understand usage behavior and optimize energy efficiency programs.

---

Dataset Summary

Column Name	Description
Household_ID	Unique identifier for each household (nominal categorical)
Date	Date of the energy usage record (datetime)
Energy_Consumption_kWh	Total daily energy consumed in kWh (continuous numerical)
Household_Size	Number of residents in the household (discrete numerical)
Avg_Temperature_C	Average daily temperature in °C (continuous numerical)
Has_AC	Indicates if the household owns air conditioning (Yes/No)
Peak_Hours_Usage_kWh	Energy consumed during peak hours (continuous numerical)

---

Key Features & Engineering

• Datetime breakdown: Month, weekday, weekend indicators
• Temperature binning: Cold (<10°C), Moderate (10–20°C), Hot (>20°C)
• Efficiency metrics:
  • Energy_per_Person = energy / household size
  • Peak_to_Total_Ratio = peak usage / total usage
• Outlier detection using Z-score
• AC Ownership binary encoding for correlation analysis

---

Exploratory Data Analysis (EDA)

Time Series Insights

• Clear seasonal trends observed in consumption.
• Weekends generally exhibit slightly different usage patterns than weekdays.

AC Ownership Impact

• Households with AC consistently consume more energy, especially on hotter days.
• AC households show a significantly higher peak hour ratio.

Temperature Influence

• Energy usage increases with temperature in homes with AC.
• Hot days exhibit higher variability and extremes in energy consumption.

Household Size Analysis

• Larger households consume more total energy but less per person (economy of scale).
• Peak consumption does not grow linearly with size.

Outlier Analysis

• Z-score based flagging reveals ~0.7% of daily entries as unusually high energy usage days.
• These could represent system faults, guests, or seasonal appliance use.

---

Visualizations

• Boxplots, scatter plots, violin plots, and bar charts created using matplotlib and seaborn
• Topics visualized include:
  • Energy vs Temperature
  • Usage by AC ownership
  • Weekday vs weekend consumption
  • Outlier trends
  • Efficiency per person

Visuals are stored in the /visualizations directory and suitable for client presentations or reporting.

---

Deliverables

File	Description
Energy_Consumption_Analysis.ipynb	Full Jupyter notebook with code, plots, insights
cleaned_dataset.csv	Cleaned dataset ready for analysis
visualizations/	Saved PNG charts from EDA
README.md	This project overview and documentation

---

Tools Used

• Language: Python 3
• Libraries: Pandas, Matplotlib, Seaborn, NumPy
• Notebook: Jupyter

---

Suggested Next Steps

• Deploy a predictive model (e.g. XGBoost or Random Forest) to forecast energy use.
• Segment households using unsupervised learning (e.g. K-Means).
• Integrate external weather data for deeper modeling.
• Build a dashboard using Plotly Dash or Streamlit for client delivery.

---

Contact

Interested in similar analyses or custom energy models?
Let’s collaborate! Reach out through Upwork or email.