We collect data from over 70 different data providers at the rate of 300 datapoints per seconds. This includes generation mixes, imports and exports, market prices and weather data. Those datapoints are then carefully validated and standardized in order to be comparable across regions.
Starting from generation and exchange data, we can trace back the physical origin or electricity consumed in each zone using flow-tracing algorithms. This is done in real-time, enabling us to compute the consumption mix of each zone, which takes into account not only local generation but also electricity imports. Using IPCC carbon emission factors, we compute the carbon emissions associated with the consumption of a unit of electricity in a given zone and at a given time.
We train artificial intelligence algorithms on our historical datasets in order to forecast the whole electricity system over the next 48 hours. Our algorithms also learn how the electricity system reacts to a change in local consumption. This enables us to assess the impact of installating an additional solar panel or smart charging an electric vehicle. As our algorithms extract structured knowledge, we can understand which type of power plant is affected by thoses changes, and where those impacts are the highest.
Showcases: Smart EV charging (Parker Project)