The Land Use Change demo currently uses proprietary prescriptor models. The models are public, but the technology to train them is not.
Train new prescriptor models than can be fully open source. Some examples can be found on the Pandemic Resilience XPRIZE challenge: https://github.com/cognizant-ai-labs/covid-xprize/tree/master/covid_xprize/examples/prescriptors