This page provides a basic tutorial.
An Example
| US Counties and Cities as of 22 Feb 2002 | ||
|---|---|---|
| n | <= n Cases | <= n Deaths |
| 0 | 375 | 1890 |
| 1 | 616 | 2284 |
| 5 | 1150 | 2690 |
| 10 | 1503 | 2852 |
| Total counties and cities - 3149 | ||
A similar procedure can be used to determine cases per day - 2999 counties have had fewer than 5 new cases in a single day, and only 8 have had more than 100.
Selected Sites
When the application begins, all the sites are selected so that a lot of functions will work right away. However, the user has the ability to select, and deselect, sites using various tools.
The most obvious are the Select all and Clear All buttons. These are both below the map and duplicated on the Filters tab. When either Clear All button is pressed, the display mode (normally controlled via the Color Keys tab) is automatically changed to Selections .
There are 4 options below the map - Add, Remove, Single site, Region - that control how the mouse affects the selections. Holding the shift key while clicking reverses the Add/Remove action.
Selections and deselections can also be made using the Search by Name controls.
Search parameters in the url can control the selections when the application starts.
Filters
By using the Range, you can select a specific day, week, or month. These values are indices (because they are easier to implement using existing components) and are displayed as days (in another field) as the values change. The default minimum is 15 Feb 2020 (index 24) for no particular reason, the default maximum is in the far future so it won't need to be adjusted for each new dataset.
You can start with everything selected and remove sites meeting some criteria, or with nothing selected and add sites.
The provided data (from Johns Hopkins) is cumulative - the numbers just get bigger (except for when they don't - see the next section). However, the application can search (and display) both the cumulative and the daily values.
There are 181 global sites (26 US states, 2003 US counties) where the slope between April 4 and 22. (80 and 91)
is less than or equal to zero.
This is good - it shows that we are recovering.
To get this, use daily slope is <= 0 with a range of 80 to 91.
If a range maximum of 200 is used (the default), the results will change everyday as new data becomes available.
Negative Daily Counts
Details
| US Counties and Cities with Negative Counts before 22 Feb 2002 | ||
|---|---|---|
| State | Anomaly | Comment |
| Washington | -512 | This was fixed the next day |
| Utah | -118 | This series has a lot of noise |
| New Hampshire | -63 | |
| South Carolina | -7 | The next day had 322 new cases |
| Total counties and cities - 3149 | ||
The application computes these values by summing the values provided for individual counties. If for some reason a county reports a lower value than the day before, then these anomalous values will appear. Sometimes, missing data is reported as zero. Using the filter function on 04-23-20, there were 880 (out of 3104) counties that had at least one negative daily value. Obviously, the reporting criteria for a (supposedly) monotonically increasing quantity varies quite a lot. (As of 04-25-20, there were 990.)
Author: Robert Clemenzi