This is their comparison point for actual radiologists. Citation number 6. It doesn't look comparable, though. Radiologists are around 90% specificity and sensitivity, which varies a good amount from the model's 77.3% and 87%, respectively.
While some data science tools may be `excel-implementable` (or some equivalent) in five to ten years, there's significant risk of mis-implementation of results given how much of a mystery many methods are to the people who'd use them.
I'd compare data scientists more to CPAs. You can have software like TurboTax and Quickbooks, but CPAs don't seem to be going anywhere. Similarly, anything that's more complicated than cookie-cutter data analysis will require someone who knows how to develop, build, and debug the algorithms themselves.
Use cases of data science are often too specific for most data science to turn into button pushers. Look at the vast array of ways a neural network can be implemented. Which one of those implementations will be in the `excel package`?
This is their comparison point for actual radiologists. Citation number 6. It doesn't look comparable, though. Radiologists are around 90% specificity and sensitivity, which varies a good amount from the model's 77.3% and 87%, respectively.