Predico - Collaborative Forecasting Platform¶
Overview¶
Welcome to Predico, a Collaborative Forecasting Platform initiative designed to enhance forecasting accuracy through collaboration between market makers and forecasters.
This documentation is designed to help forecasters to learn more about the platform and interact with it, effectively.
Important
This platform is a proof of concept, so there might be changes to improve the outcomes for all parties.
This platform was developed and is maintained by INESC TEC Centre for Power and Energy Systems (CPES) developers.
Which forecasting use cases are currently available on Predico?¶
The Predico platform currently creates forecasting challenges for two use cases:
- Forecast the day-ahead Offshore Wind Power Generation in the Belgian grid.
- Forecast the day-ahead Solar Power Generation in the Belgian grid.
These challenges are open for forecasters to submit their forecasts daily. The forecast horizon is day-ahead, and new challenges are created daily, per market session. You can find more information about market sessions, challenges and the remaining nomenclature used within the platform at the Introduction section.
How can I interact with the Predico platform?¶
There are two ways to interact with the Predico platform:
- Web Application: This is the official web application of Predico, where you can visualize your forecasts and their scores.
- REST API Swagger: This is the official REST API of Predico, where you can interact with the platform programmatically.
- The API is designed to be used by forecasters to submit their forecasts and retrieve their scores.
- The API is also designed to be used by developers to create their own applications and tools to interact with the Predico platform.
Where to start?¶
Start by the Introduction section to understand the critical definitions and the general workflow of our collaborative forecasting platform.
Want to know more about Predico ensemble approaches? Visit Our forecasting models section.
Documentation Outline¶
This platform consists exclusively of an API server, with no front-end available. Therefore, every interaction is done via HTTPs requests.
The following sections are designed to be followed sequentially to ensure a smooth interaction with the Predico API Platform.
- Authentication: Learn how to authenticate with the Predico API Platform.
- Listing Open Sessions: Retrieve details on the open market sessions.
- Listing Challenges: Retrieve details on the challenges published by the market maker, during open market sessions.
- Downloading Raw Data: Download raw measurements data for a specific challenge.
- Preparing a Session Forecast: Prepare a forecast submission for a specific market session, based on the retrieved data.
- Submitting a Session Forecast: Submit the session forecast to the Predico API Platform.
- Preparing a Continuous Forecast: Prepare a continuous forecast based on the retrieved data.
- Submitting a Continuous Forecast: Submit the continuous forecast to the Predico API Platform.
- Listing Continuous Forecasts: Retrieve the continuous forecasts you have submitted to the Predico API Platform.
- Listing Session Submissions: Retrieve a list of forecasts submitted by the forecaster.
- Listing Session Submission Scores: Retrieve the forecast skill scores for your submissions.
Important
While the provided code examples are demonstrated in Python, the fundamental logic and sequential processes are universally applicable. Forecasters are encouraged to develop their own clients using any programming language of their choice, adapting the concepts and workflows presented here to suit their preferred development environment.