“We’ve learned that the best data scientists are skeptics and follow Twyman’s law: Any figure that looks interesting or different is usually wrong. Surprising results should be replicated—both to make ...