Turning an Idea into a Data-Driven Production System: An Energy Load Forecasting Case Study

We solve a wide range of problems in our daily lives that do not require excessive thinking, instead relying on data from past experiences. Somewhere in our brain we use data-driven models that help us make decisions in different situations. It is our own data-driven production system.

Likewise, at a business level, you may use data analytics to turn large volumes of complex raw data into actionable information that can improve your engineering design and decision-making processes. However, developing such effective analytics and integrating them into business systems can be a challenging process.

Data is being collected today at a pace we couldn’t have imagined years back. Individuals, companies, organizations, and governments are collecting data from everywhere: from sensor data coming from your DIY weather station to images captured by traffic cameras in the city center.

Building a prototype and validating an idea using the available data might be the very first step before scaling up the problem to a full production system. The vast amounts of data available today have made it possible to create highly accurate forecast models. The challenge lies in developing data analytics workflows that can turn this raw data into actionable insights. A typical workflow involves four steps, each of which brings its own challenges:

1. Importing data from disparate sources such as web archives, databases, spreadsheets, etc.
2. Cleaning the data, identifying and removing outliers, and syncing up the data
3. Developing an accurate predictive model based on the aggregated data using machine learning techniques
4. Deploying the model as a scalable application in a production environment

In this session, using an energy load forecasting case study for the State of New York, you will see how MATLAB^® can be used to complete the entire data analytics workflow, turning an idea can into a data-driven production system that uses an accurate machine learning model.

Energy producers, grid operators, and traders have to continuously make decisions based on an estimate of future load on the electrical grid. As a result, accurate forecasts of energy load are both a necessity and a business advantage. Traders might determine an appropriate trading strategy for a given day, and energy producers or grid operators could use the results to understand the effect of weather on energy loads and determine how much power to generate or purchase. Given that the State of New York alone consumes several billions of dollars of electricity per year, the result can be significant for power generation companies.