While I *did* attend Stanford, I’m not just highlighting their work for a second day in a row due to personal bias.
For modelling derivation or validation applied to emergency care in a modern health system, this may be the most important data set ever publicly (to credentialed users who sign the data use agreement) released.
Over 100,000 ED visits with all this granular data – including *waveform* data from physiologic monitoring:
This is a treasure trove of data, far richer than any other currently publicly available. For less-sophisticated health systems, including those around the world, access to these data go a long way towards democratizing data science in emergency care.