What if our data is scattered or unlabelled?

“It's the most common situation on the plant floor, and it's part of the initial diagnosis work. Before modelling anything, we map the data's sources, quality, frequency and semantics.”