Text mining for forecasting the development of electricity prices

Challenge

In order to raise efficiency potentials and open up new business areas, an energy provider is increasingly relying on data mining and predictive analytics. In order to support the company’s decisions in a targeted and reliable manner as much and as varied data as possible is to be analyzed . In concrete terms, for example, the news from the Reuters news agency is recorded every minute for electricity trading and individual news items are examined for their relevance to the development of energy prices.

Solution

The company has decided to implement the necessary algorithms itself and to optimize them in the future, as the know-how around data analysis is considered critical to the company. The solution is to be implemented with the free statistics language R. The main arguments in favor of R are the wide range of possibilities in all areas of data mining and predictive analytics, the high quality of the software, and the integration options into the existing infrastructure. In addition, there are no licensing costs for R.

With the R-Academy, eoda offers a comprehensive training program for R, which can be specifically expanded in individual training sessions.

Against the background of the customer’s goals and with eoda’s experience, a tailor-made training offer was designed to provide the participants with the greatest possible benefit. The concept thus created had two basic focal points:

• introduction to data analysis with R.
• text Mining with R

The second part of the training covered the analysis and classification of texts with R. Reuters messages were prepared using tokenization, sentence splitting and normalization, among others, and then analyzed using methods such as support vector machines and neural networks.

Result

Like many other companies, the energy provider has recognized the potential of the free programming language R for data mining and predictive analytics. R can be integrated optimally into the existing infrastructure and offers the opportunity to implement the highest-quality and most modern analysis methods in a timely manner.

eoda’s individual training courses provide customers with significant support in developing their own analysis solution and using it productively. The analysis of thousands of messages per day with regard to company-specific goals is a typical example of the requirements in the big data environment. eoda’s training courses provide the knowledge that the responsible employees at the customer’s site need in order to use “big data” profitably for their own company.

eoda also supports the energy supplier in the follow-up to the training with its R-support and methodological advice on the content of the training. This ensures that the achievement of the predefined goals is optimally supported even after the end of the training and that the benefits from the use of R are maximized.

YUNA delphi – for better forecasts

From purchasing to sales, we optimize the quality of your forecasts with YUNA delphi. YUNA delphi brings together your expert knowledge and our reliable forecast models in a unique approach.

 

Learn more