Small and medium-sized retail suppliers need a reliable view of their portfolio position. When generation and consumption change throughout the day, simple planning quickly turns into expensive imbalance.
Predicta combines historical consumption and generation data with market and weather information. The result is a quarter-hourly forecast that can be tested safely through paper trading before it is used in live operations.
1. Anonymised data
Some customers cannot share EANs directly. Each consumption or generation point is therefore assigned an anonymised ID. The data architecture supports regular delivery of historical and daily data within the agreed NDA framework.
2. Historical data analysis
High-quality input data is the foundation of a reliable model. Every customer portfolio has unique characteristics shaped by its consumption and generation points, customer types, energy sharing and other factors. Each portfolio therefore requires its own modelling approach.
The analysis covers seasonality, trends, autocorrelation and relationships with external variables. Model selection, external data integration and parameter optimisation follow from those findings. Useful external inputs can include weather forecasts, calendar effects, price signals and regulatory interventions.
Once the results reach an acceptable level, the process moves to paper trading.
3. Paper trading: forecasts and shadow trading
At this stage, the customer receives an alternative forecast of its position based on OTE data. The open position is then covered through shadow use of the spot markets, including the day-ahead market, intraday auctions and the continuous intraday market.
The resulting trading strategy can be evaluated against the portfolio’s imbalance costs. If the results remain satisfactory after a period of calm review, cooperation can move into live operation.
4. From forecast to daily operation
To avoid daily Excel downloads and unsafe copy-and-paste workflows, Wattflow provides a dedicated operational tool. It supports automatic forecast refreshes every three hours, a current view of the balance responsible party’s position, position shifts from day-ahead to intraday markets, order scheduling, delivery-schedule management, trade monitoring and reconciliation.
Once the results are understood, the forecasts can support day-ahead planning, intraday position updates and automated balancing in Wattflow. Models are monitored and improved as new data becomes available.
If your imbalance costs could be lower, or your portfolio positions could be managed more safely and precisely, the situation can be discussed without obligation. Contact details are available at www.predicta.cz.