Not bs. But there is both a real GIGO problem, and a problem with under specification. It's certainly easy to propose DS analysis that are unlikely to have mu

Not bs. But there is both a real GIGO problem, and a problem with under specification. It's certainly easy to propose DS analysis that are unlikely to have much return. Thinking "data science is hot, we should do that" is different than "we have all this data and don't understand what it means". The latter is more likely to lead somewhere interesting.