Back to work

Energy Demand Analysis & Visualization

2025 · Data Science

Energy Demand Analysis & Visualization

Forecasting New Zealand's energy demand and emissions with time-series models.

Role

Independent Project

Industry

Data Science

Year

2025

Overview

An independent data-science project forecasting New Zealand's residential energy consumption and carbon emissions. I developed and tuned ARIMA and SARIMA time-series models, incorporating GDP and population as exogenous regressors to improve accuracy, and rigorously validated them with MAE, RMSE, and MAPE. The forecasts are served through a lightweight Flask API and brought to life with interactive Power BI and Python visualisations.

The challenge

Energy demand is driven by many interacting factors. Producing forecasts accurate and explainable enough to support real sustainability insights is genuinely hard.

My approach

I built and tuned ARIMA/SARIMA models with GDP and population as exogenous regressors, validated them with MAE/RMSE/MAPE, and exposed the results through a Flask API and clear visualisations.

The outcome

The project produced reliable forecasts and interactive dashboards that turn raw consumption data into trends anyone can explore — supporting energy-efficiency and sustainability decisions.

What I did

  • ARIMA / SARIMA time-series forecasting
  • GDP & population as exogenous regressors
  • Accuracy measured with MAE, RMSE, MAPE
  • Flask API serving live forecast data
  • Power BI + Python (Matplotlib/Seaborn) dashboards

Built with

  • Python
  • ARIMA/SARIMA
  • Flask
  • Power BI

Want something like this?

Tell me about your project and I'll send a clear, no-obligation quote.

Start a conversation