Skip to content

Products

DataFlow, DemandCast and MetaCast

DataFlow collects the data and delivers the forecasts. DemandCast produces meter-level consumption forecasts. MetaCast optimises either the DemandCast forecast or the supplier's own forecast, taking the system direction into account.

  1. Data sources EPİAŞ · OSOS · OSB · Weather
  2. Data layer DataFlow
  3. Forecast DemandCast or your own forecast + Forecast Optimisation
  4. Optimisation Imbalance Optimisation
  5. Delivery EPİAŞ · Trading platform · Your team

Data and integration layer

DataFlow

DataFlow collects, cleans and prepares the data the models need, and delivers the forecasts to where they are used.

Included with the forecasting products.

  • Data collection

    Settled meter consumption and the current portfolio meter list are pulled from EPİAŞ automatically every month. Hourly consumption from OSOS and weather data are collected continuously. For meters inside organised industrial zones (OSB) that you cannot reach through OSOS, data is pulled from the OSB portals.

  • Data quality

    Missing data is filled in and faulty data is removed. Where there is no history, as with new meters, synthetic data is generated with AI.

  • Delivery and integration

    Forecasts are submitted to EPİAŞ automatically and passed to trading platforms such as SmartPulse and Senkron in near real time.

  • Reporting

    Daily forecast and OSOS consumption reports arrive by email as Excel or CSV files. Data from different distribution regions is available in one place.

  • Dashboards

    Forecast, actual consumption, imbalance volume and cost are shown on the same screen. Performance is tracked monthly across the portfolio and daily for OSOS meters.

  • Optional module

    KGÜP and EAK automation

    Automates the KGÜP (finalised day-ahead generation schedule) and EAK (available capacity) notifications for suppliers serving OSBs. OSBs can also use it directly.

Consumption forecasting

DemandCast

Each meter in your portfolio is modelled separately, so differences in consumption behaviour feed directly into the forecast. The whole portfolio is covered, including meters without OSOS data.

−22.7%Imbalance costAgainst the supplier's own day-ahead forecast, average
−22.1%With intraday revisionsFurther reduction in the remaining cost*

2026, live operation. *The intraday effect depends on trading strategy and algorithmic trading rules.

  • Day-ahead forecast

    The next day's hourly forecast is produced every day in line with the day-ahead (GÖP) submission schedule. It is sent to you, to your trading platform or directly to EPİAŞ.

  • Intraday revision

    Once the intraday market (GİP) opens, the forecast for each contract is revised every hour with the latest data. Revisions continue until the contract closes; the timing of submissions is set to your needs.

  • Submission rules

    Which revision is sent, and when, is set by the supplier's rules. Trading decisions stay with the supplier.

Optimisation layer

MetaCast

Runs on top of the forecast and takes the system direction into account, that is, whether the system is in surplus (long) or in deficit (short) in a given hour. It has two modules that can be used together or separately.

−16.7%Imbalance OptimisationFurther reduction in the remaining cost
85.5%System direction accuracyIn hours where the model is at least 80% confident

Imbalance Optimisation: 2026, live operation. System direction accuracy: January–August 2026, 5,390 hours; 68.8% across all hours.

Forecast Optimisation

Corrects the forecasts you already use, whether they come from your own team or another provider, by learning from their past errors.

Works on day-ahead and intraday forecasts. Your workflow does not change; the corrected forecast is sent to you.

Imbalance Optimisation

Predicts the system direction and adjusts the final forecast to minimise imbalance cost. It applies both to DemandCast forecasts and to forecasts that have been through Forecast Optimisation.

  • Day-ahead · in use
  • Intraday · on the roadmap

All modules together

−47.2%

In a portfolio using meter-level forecasting, Imbalance Optimisation and intraday revisions together, imbalance cost was brought down from 48 to 26 TL per MWh consumed over six months.

Results step by step →

Packages

The day-ahead forecast is the base; system direction and intraday revisions can be added

Base product Day-ahead Day-ahead + Intraday Day-ahead + Direction Day-ahead + Direction + Intraday
DemandCast ✓✓✓✓
MetaCast Forecast Optimisation ✓✓✓✓

"Direction" refers to Imbalance Optimisation. DataFlow is included in all packages. Pricing is set in a proposal based on portfolio size and modules.

We calculate the potential contribution of each module to your portfolio separately, through a backtest

Request a meeting