Summary
DataRobot provides us with a unique capability to rapidly develop models. With this platform, data scientists can combine the benefits of DataRobot and the flexibilities of open programming. In this chapter, we explored ways to access the credentials needed to programmatically use DataRobot. Using the Python client, we demonstrated ways in which data can be ingested and how basic projects can be created. We started building models for more complex problems. We created model factories as well as one versus all models. Finally, we demonstrated how models can be deployed and used to score data.
One of the key advantages of programmatically using DataRobot is the ability to ingest data from numerous sources, score them, and store them in the relevant sources. This makes it possible to carry out end-to-end dataset scoring. It becomes possible for a system to be set up to score models periodically. With this comes numerous data quality and model monitoring concerns. The next chapter...
