This is accurate. The gulf between scientists' actual digital struggles (bad calibration, data turf wars, broken formats, proprietary software, poor support & infrastructure, no training) and management or external hopes / dreams / perception is big.
Stats and machine learning people are often disillusioned as they discover our datasets aren't THAT big, and have weird patterns of error and missing data as a result of hailstorms, broken equipment, and global pandemics.
#healthcare version of What #technologists think #HCP need: Would you like virtual-reality machine learning remote real-time robot nanosurgery teaching? What #HCP actually need: Could you give me mobile phone network coverage in the hospital please?
Scientist Tech Help First I need the metadata from the camera.... :p
TIL about WebPlotDigitizer <. On the one hand, it’s great that this exists! On the other hand, like Randall Monroe < I also deeply resent the need for its existence.