Fully agree! In the case that tabular data means business process records, there's absolutely no need to use anything more complicated than a principled statistical model. Focus on the quality of the data and the business problem, not exotic ML.
In more detail: the business data is prone to: evolving processes / systems / products / markets / customers; errors / omissions / corrections; tail events; hirings / firings; data loss; etc etc. The datasets tend to be small, messy, complicated, subjective. Nothing about this suggests needing a large, complicated model.
I'd be surprised if geographical proximity to a BLS4 made much a of a difference to what would otherwise be a rudimentary swab & routine inspection (if that was even done at all).
As Kristian Andersen notes, the background noise of flu season would surely drown out the weak signal of an unknown novel virus. Then to not only notice the weak signal but act on it so quickly to do primary research and characterise it as a novel virus within such a short timeframe?
Seems far less likely than simply the effect of poor operational standards.
That comment on that page from Kristian G. Andersen is mildly disturbing (http://archive.is/O1vhN)... As Associate Professor, Scripps Research, Director of Infectious Disease Genomics, SRTI etc etc I assume he knows what he's talking about when he states:
"That means that the outbreak was detected almost immediately after the first case, which - given that this is flu season in China - is just amazing. Detecting an outbreak of pneumonia (similar to flu) of a novel coronavirus that fast is truly impressive."