Delta – how different is your dataset?

Now that you have some idea of what use case you are most interested in, and what datasets will give your organization the most value, it’s time to understand how unique your dataset is. This analysis matters because it will answer two questions:

1. First, which models are already on the table for you to use, due to having been trained on similar data?
2. Second, how well have those models performed?

This insight will start to give you a clue toward what performance you can hope to achieve on your datasets as a best-case scenario. Then, we’ll plug that expected performance number back into our total project value and make sure we’re still on track. The next chapter is completely dedicated to answering those questions. Here, we’ll learn how to pick apart your dataset. This is a good section for those who are new to data analysis.

First, it’s always a good idea to spend time really analyzing...