I'm a scientist in industry. It's remarkable how many smart people think that science can be done without data. I've heard managers ask: "Why do we need to gather data? Can't we just model it? The customer doesn't want to see data. They just want an answer."

There's also a strong belief in "statistical magic." Faced with a bad or insufficient data set, someone will say: "Let's give the data to <statistician> and have them work their magic on it."

That the results actually have to be influenced by the data in some way is something that has to be explained to people. In all of my years as a scientist, I've learned that there's still no substitute for good measurements. Good data can be cheaper than analysis of bad data.

Data science is not fundamentally about data or science. It's about either justifying decisions that have already been made or delegating decisions to an unbiased casting of bones to let the gods decide.

> It's remarkable how many smart people think that science can be done without data.

It’s so important that we write these down, so when these people have forgotten why they’re not making any progress and they’re searching for answers, they’ll find what we wrote down and say “ohhh, we had too much hubris thought we were smarter than everyone else and didn’t listen to how important actually going outside is.”