Good morning Mission. And welcome to Virus Village, your daily data dump.
Not a bad picture for an otherwise pretty nice morning in SF. Following the pattern of the past few days, the total number of positive cases continues to rise, but the trend line over the last few days is not terrible. No new deaths have been recorded. The R number estimate remains stable, but well over the goal, which does not augur well for numbers in the coming week.
Scroll down for more.
HiGeorge, a data visualization startup, developed some new visualizations for Mission Local, which we will be using and fine-tuning in the days to come.
When we say there are 108 new cases, what we mean is that 108 cases have been added to the City totals, not that there were 108 new cases yesterday. Those 108 cases come from earlier tests which, after delays, were finally analyzed and reported. Given the way the city reports its numbers, it’s best to look at the seven-day average.
The Mission District reported 9 new cases, for a total of 858.
Note the graph above, showing the rolling 7-day average of cases per day for the week ending July 22 to be 64.
According to DPH, the rolling average, is “the average of new cases for a particular day and the prior six days which shows the trend (smoothing out daily fluctuations).”
Looking at the rolling average gives a different, more positive picture.
The “key indicator” the City uses is the weekly average number of new cases per day per 100,000 residents. Their goal is an average of 1.8 new cases. On July 21, that number was 9.3 (equivalent to an average of 81 new cases a day). So the rolling average may be falling, but it is still quite high (in the City’s red zone).
Like the new case figures, the positivity rate (the proportion of positive results of daily testing) has fluctuated over the past week. The fluctuation and the ambiguity is a result of the “high degree of variation in the time needed to compete tests by different labs.” The figure for July 22, based on 1,183 tests is 4.3 percent, higher than the overall average of 3 percent since March.
Hospitalization figures have apparently not changed from yesterday with 122 COVID patients counted as of July 21. The number includes confirmed and suspected COVID patients as well as transfers.
Note the number of COVID ICU patients in the graph above represents a rise of 4 patients from July 20.
One of DPH’s key indicators is the rate of weekly change in COVID patients. The week ending July 22 showed an increase of 17 percent, well above the targetted weekly average of 10 percent.
The seven-day average percentage of Acute Care bed availability for the week ending July 22, is 21 percent. This percentage has steadily fallen since July 7.
This average percentage of ICU bed availability is 29 percent as of July 22. This average percentage has fluctuated since June 29.
Both Acute and ICU availability figures are greater the DPH target of 15 percent availability.
Men have contracted 56.4 percent of the positive COVID cases and 66 percent of the deaths recorded. Those who are unhoused have accounted for 230 cases and one death. Community contact (as distinguished from contact with a known case) continues to be the largest source of transmission for cases and deaths.
It should be noted that the data in your graph used to calculate the rolling average for the last 7 days is based on positive results for tests given on a specific day. Due to the delay to get test results, the graph of new daily usually shows a drop in the rolling average line in the last 7 days. Over the next 7 days, the positive numbers for the last 7 days are adjusted upwards as more test results come back from the labs. So the rolling average number on Jul 22 of 64 should not be directly compared to the rolling average number on Jul 15 of 109 on the SAME graph. In the data set from 7 days ago when the last day was Jul 15, the rolling average for Jul 15 was 80, not 109. So for a more apples-to-apples comparison, the rolling average did drop, but from 80 to 64. not 109 to 64.
Putting this another way – the data set from 7 days ago shows the rolling average for Jul 15 as 80. The data set from today shows the rolling average for Jul 15 as 109.
Another clarification is the graph comparing deaths for the 6 Bay Area counties is for totals without regards to the county population. If the deaths are shown per 100,000 population, Marin and San Mateo are about 50% higher than Santa Clara and Alameda.
Thanks. I very much appreciate your comments and I understand what you are saying which makes sense. As this is far from my field, I have been relying on the daily information I get on the SF Open Data site from the DPH which obscures the point you are making. I hope to talk with DPH, and the “rolling seven day average”, how to present it and interpret it is one of the questions I hope to get explore with them. Your point on the deaths per 100,000 for the graph also makes sense. Any other comments, suggestions, or corrections will be gratefully appreciated.
It still would be helpful to see dates on the rolling average graph, they are either covered by the donate button or not there!
Thanks Lester. I think the best way to get the date on that graph is to move your cursor over the line. It is interactive. If you have any other problems, or is this is not what you meant, please let me know.