[burger]

Im sure this is out there. Just want to find it- ratios of positives to number of tests over time.  Calif is hitting record positives. Also record number of tests. One would think positive to test rate is an important stat.  

The only comment I saw relating to Calif’s Tuesday’s record number is that the percentage of positive per tests  is “slightly up.”  

Isn’t that an important number to track?  Is there a site I can track it on?

Here's another tracker:

https://covidactnow.org/us/ca?s=56070

The dashboard is useful:

The estimated infection rate ® is above 1, but not by much.
The positive test rate is 5%.  That's up from a trough of 3.8%.
ICU headroom used is up to 29%.  That's plenty of room to handle a surge, but is up from 20% earlier.
Contact tracing is down from 30% to 21%, not a good trend.

BC

These trackers that have pre-defined criteria for when things are good or bad concern me (the state uses similar metrics).  For one thing, these mostly focus on levels, not rates of change.  ICU % usage may be in the good range now, but if hospitalizations are increasing rapidly, that won't last for long.

Also, I would love to know how they're calculating R.  If we're setting records for cases every day and the % of positive tests is increasing, I have trouble imagining how each case in California is only resulting in 1.05 new cases.  I think that estimating R from these data is extremely difficult.  The number of tests conducted is changing from day to day, so you have to account for that when estimating R.  That raises another complication: what's the expectation for % positive as a function of the number of tests given?  I have searched around for a discussion of this issue and found nothing so far.  I'm sure that this can be modeled, but it would take a complicated model that incorporated the probability of getting tested as a function of the number of tests given and whether or not someone is infected.  That would be a mess. 

Ultimately, this does not have to be complicated.  My dashboard would only have two questions: 1) is the number of cases going up faster than the number of tests?  2) Are hospitalizations increasing?   If yes to both, you need to do more to get the virus under control.

[burger]

Im sure this is out there. Just want to find it- ratios of positives to number of tests over time.  Calif is hitting record positives. Also record number of tests. One would think positive to test rate is an important stat.  

The only comment I saw relating to Calif’s Tuesday’s record number is that the percentage of positive per tests  is “slightly up.”  

Isn’t that an important number to track?  Is there a site I can track it on?

Here's another tracker:

https://covidactnow.org/us/ca?s=56070

The dashboard is useful:

The estimated infection rate ® is above 1, but not by much.
The positive test rate is 5%.  That's up from a trough of 3.8%.
ICU headroom used is up to 29%.  That's plenty of room to handle a surge, but is up from 20% earlier.
Contact tracing is down from 30% to 21%, not a good trend.

BC

These trackers that have pre-defined criteria for when things are good or bad concern me (the state uses similar metrics).  For one thing, these mostly focus on levels, not rates of change.  ICU % usage may be in the good range now, but if hospitalizations are increasing rapidly, that won't last for long.

Also, I would love to know how they're calculating R.  If we're setting records for cases every day and the % of positive tests is increasing, I have trouble imagining how each case in California is only resulting in 1.05 new cases.  I think that estimating R from these data is extremely difficult.  The number of tests conducted is changing from day to day, so you have to account for that when estimating R.  That raises another complication: what's the expectation for % positive as a function of the number of tests given?  I have searched around for a discussion of this issue and found nothing so far.  I'm sure that this can be modeled, but it would take a complicated model that incorporated the probability of getting tested as a function of the number of tests given and whether or not someone is infected.  That would be a mess. 

Ultimately, this does not have to be complicated.  My dashboard would only have two questions: 1) is the number of cases going up faster than the number of tests?  2) Are hospitalizations increasing?   If yes to both, you need to do more to get the virus under control.