These graphs work best on a computer. Phones, not so well. Phone users: check out the blog.

These graphs are interactive (and not very useful or intelligible if viewed without interacting). You can click-and-drag to zoom, single-click on a particular item to turn it on or off, or double-click to show just that item.

### Severity

The following graph attempts to show broadly “how bad” the epidemic is in each state, based on daily cases. It does this by multiplying-daily-cases-per million (DCPM), by the week-over-week growth factor in new cases (DCPM and growth are graphed separately farther below). Over 100 is a “severity warmspot” (bad); over 200 is a “severity hotspot” (very bad). Here are the top 20 states:

And for all states:

### Total Cases

### Daily Cases

##### Daily Cases Per Million People

To really compare the numbers of different regions, we need to adjust for population. Hence “daily cases per million” (DCPM), which lets us see how *relatively* bad a region has it. I define a region as “warm” if DCMP is more than 100, and “hot” if DCPM is more than 200. You’ll definitely want to click-and-drag to zoom in to the busy part of this graph, since there are a few wild outliers that smash everything to the bottom:

The graph below is similar but shows all states, and shows the moving average rather than the individual days with trendline (top 20 states):

And for all states:

###### New York Times Data (Cases)

Recently (December 2020), the coronadatascraper.com data seems to have been getting worse, e.g. no data for North Dakota or recently California. So here’s the same data from the New York Times data collectors; I may shift away from coronadatascraper.com and to this in the future, for daily cases at least:

And for all states (still New York Times data):

##### Daily Cases Trajectories

The next graph shows the relative trajectories in daily-cases-per-million of the states, in days since each state reached 100 DCPM (those that haven’t are not shown). So each state’s graph is *shifted* in time, so they align horizontally at the moment they reached 100 DCPM:

##### Daily Cases Growth

This next graph is the “rope chart,” which shows the daily growth of new cases each region. This allows to predict the direction a region is headed. When a line crosses zero headed downward, that region has turned the corner toward recovery. I define a region as “warm” if growth is greater than 1%, and “hot” if growth is greater than 5%. Remember that you can double-click a name at the right, to see a graph of just that region. Here again you’ll probably want to zoom in with click-and-drag to see the details better:

This next graph is similar to the rope chart, in that it shows percent growth. But rather than fitting a broad trendline, it uses a smoothed moving average, to show more of the fine details of the trajectories of each state.

And the same for all states:

###### New York Times Data (Growth)

Recently (December 2020), the coronadatascraper.com data seems to have been getting worse, e.g. no data for North Dakota or recently California. So here’s the same data from the New York Times data collectors; I may shift away from coronadatascraper.com and to this in the future, for daily cases at least:

And for all states (still New York Times data):

### Hospitalizations

Hospitalizations per million, for all states, as a 7-day moving average:

### Deaths

##### Daily Deaths Per Million People

### Testing

##### Percent Tested

##### Percent Positive

This measures how many tests come back positive. For comparison, a positive test percentage of less than 8%, is considered good enough to reopen, by the state of California.

Here are the positive test rates, for all states, shown as a 7-day moving average, which shows the short-term trends much more than the broad trendline above. Less than 8% is doing OK; less than 2% is very good: