The prediction of future biodiesel demand in different regions of the world involves an extensive data collection process. A prediction model based on this data must not only consider available information on historical fuel consumption and related quantities, but also consider current and future constraints imposed by local biofuel mandates. Evonik is interested in a computational model that can predict future biodiesel demand per geographical region (e.g. on the country level) for a time horizon of 10 – 15 years. This will help Evonik in assessing future demands and opportunities for its biofuel catalyst business and how it contributes to corresponding sustainability criteria.
Historical data on biodiesel consumption and information on regulatory constraints are currently collected and aggregated manually from various data sources. These sources include publicly available data from governmental websites and national statistical services, as well as additional data from commercial service providers.
This use case aims at developing an automated data collection process that aggregates publicly available information about biodiesel consumption and regulatory constraints (biofuel mandates) in a centralized data repository. The process should provide the possibility for regular automatic data collection and updating (e.g. on a daily or weekly basis). A coherent data structure must be set up to integrate the collected data with additional internal data sources. In conjunction with expert domain knowledge available at Evonik, this structure shall subsequently serve as the basis for training and evaluating the required prediction models. Within the scope of this challenge, also the development of a prototypical prediction model is desired, which utilizes the aggregated data sources and predicts the expected biodiesel consumption for an exemplary geographical region.
If the results of this use case are satisfying, a possible follow-up project will extend the solution to worldwide market coverage and additional data sources to improve the prediction accuracy even further. More specifically the data will enable the development of scenarios with respect to global greenhouse gas savings scenarios and enable to choose favorable business conditions with the goal of maximizing these savings.
If you have applied and you want to participate at the Q&A - Session please register by clicking this Link.