Weather-Driven Power Demand Forecasting Using API Integration and Predictive Modeling

The Challenge

Electricity demand in Vestmannaeyjar is highly weather-dependent, as power is primarily used for heating. However, operational planning lacked a structured way to link weather forecasts to future MW demand.

The key challenges were:

• Integrating reliable external weather data into TrendMiner
• Correlating wind, temperature, humidity, and sunshine with electricity consumption
• Building a predictive model to estimate required MW production
• Transitioning from reactive analysis to forward-looking forecasting

Without predictive capability, production planning relied on historical patterns rather than forecast-driven optimization.

‌

The Approach

A predictive workflow was built combining external weather data and TrendMiner’s calculation capabilities:

• Weather data (including “feels like” temperature and wind) was integrated via API (open-meteo) using custom calculations
• Correlation analysis was performed between weather parameters and electricity consumption
• A regression model was created using the Prediction Tag to estimate total energy demand
• The resulting equation was implemented in a formula tag
• Instead of using current weather values, forecasted weather data for the next 24 hours was applied to predict upcoming energy needs

This transformed TrendMiner from a monitoring tool into a short-term demand forecasting engine.

‌

‌

The Results

‌

‌

The Takeaway

The organization now has a predictive tag that estimates upcoming energy demand based on weather forecasts, enabling more accurate production planning, improved load balancing, and reduced operational uncertainty in weather-sensitive power generation.

‌