You might want to give http://www.gleamviz.org/simulator/ whirl. You can simulate any epidemic model on a real-life datasets of mobility and healthcare access around the world.

The tricky part is know exactly what the "correct" epidemic model is: length of latent/infectious/etc periods, etc.

Full disclosure: I was once one of the main developers of this epidemic modeling framework.

Coincidentally I was just looking at gleamviz the other day (I was googling around for accurate modeling programs). Sadly when I tried the linux version, it segfaults after some Qt error.

This is really cool!

Are there any existing models to start with?

I've been disconnected from GLEaM for a while now, so I might be missing something better, but the user manual (http://www.gleamviz.org/simulator/GLEAMviz_client_manual_v7....) guides you thorough an example on Sec 4

If containment measures begun on January 22 are effective, cases should begin falling from trend over the next few days (27th through Feb 5 or so). We're at about 10,000 identified cases, I expect that will climb to ~50k, with about 1,000 - 2,000 deaths.

I'm not an epidemiologist, but rather a space alien cat, so factor accordingly.

Not a prediction, but if what I'm reading in the news is true, it's supposed to peak within two weeks.