Q: Why doesn’t the latest data date range include today’s date?
A: The data is pulled down from the devolved government websites and sometimes there is a few days lag – especially at weekends and over bank holiday weekends – in the current day’s data becoming available.
Q: Does that mean that the situation could be worse than portrayed in the graph?
A: Potentially, but only slightly, which is why the system always pulls data from the last 10 days and provides ascertainment biases of 2 and 5 to help account for any shortfall. It could also be slightly better if the trend is for falling cases, but better to err on the side of caution.
Q: Why 10 day’s data?
A: The recommended isolation period from the Government, PHE, NHS (11-Dec-20 – see UK Covid self-isolation period reduced from 14 to 10 days) and others is to isolate for 10 days if you have a positive test, journey to the UK from one of the restricted countries or have been in contact with somebody who has tested positive.
Q: How often is the database updated?
A: I try and do it every day in the afternoon but that’s not always possible and some days the devolved Governments don’t update the data on their websites. The Northern Ireland Government, for example, doesn’t update their data over weekends or bank holidays.
The database was last updated on Latest update: 17:46, Friday 17th September, 2021.
Q: How current is the data?
A: The data from the UK Government’s database is updated nearly every day. However, the data is not fixed, it is constantly being revised; even data that is 6 months old can and does change. Number of cases data usually settles down to a fixed value after a week / 10 days and then stays more or less fixed but can fluctuate by 1 or 2 cases over time.
Q: What is the science behind the analysis performed?
A: The method originally comes from Georgia Institute of Technology, a well respected US university. This website uses exactly the same analysis method and the validation page demonstrates how the method has been transferred from the Georgia Institute of Technology website to this website.
Q: Why ascertainment biases of 2 and 5?
A: Not everybody who has the virus has symptoms – they are asymptomatic – and the Government’s Track and Trace system does not successfully Track and Trace everybody who has been in contact with somebody who has tested positive. Therefore, there needs to be a margin between those tested and those in the community, who may have the virus and may not be aware of it or who do not isolate correctly even though they know they have the virus – see Less than 20% of people in England self-isolate fully, Sage says (Guardian article dated 11-Sep-2020). The Georgia Institute uses biases of 5 and 10 but somewhere in the range of 2 and 5 seems to be about the correct bias for the UK, but it could be worse. The about page has more information. The recently added Infection Ratio number to the mouseover data on the 2 graph views shows Infection Ratio numbers similar to those being talked about on the TV – see Coronavirus (COVID-19) Infection Survey, UK: 29 January 2021 and The Guardian headline on 19-Feb-21 (The Guardian 19-Feb-2021) stated that 1 in 115 people had Covid-19 for the week ending 12th February, which equates to an ascertainment bias of approximately 4, close to the upper line on the plots in this tool. Interestingly I’ve noticed that the Georgia Institute have changed their bias numbers to 3 and 5 (mid Feb-21) from the values 5 and 10 that they used when they initially launched their tool. Random checks on the data provided by the C-19 by ZOE app (throughout 2021) show that the ascertainment bias has regularly sat in the range 4 to 5.
Q: What’s the relationship between ascertainment bias and infection ratios and how do these graphs reflect the infection ratios being used in the media?
A: There is no fixed relationship except that as the ascertainment bias increases the infection ratio increases i.e the number falls from 75:1 to 50:1. Passing a mouse over the plotting area of the graphs will display an information box with the X and Y axis data points, the bias number and the infection ratio. Currently, early Feb 2021, as the cursor approaches the ascertainment bias of 5 line you should see infection ratios displayed of around 50:1; meaning that 1 in every 50 people in the community is infected. This report from the UK Government ONS website Coronavirus (COVID-19) Infection Survey, UK: 29 January 2021 from late January 2021 mentions similar infection ratios suggesting that the asertainment bias line of 5 is the one to concentrate on when using this tool.
Q: How accurate is the analysis?
A: The analysis is only as good as the data provided. The data presented is not aimed at demonstrating the probability that you will catch the virus but the likelihood that somebody infected with the virus will be at an event. If you take the recommended additional precautions – space, mask, wash – then your chances of catching the virus reduce. The validation page explains the maths behind the analysis.
Q: What does 50% chance, for example, mean?
A: A 50% chance means that there is an odds-on even chance that somebody at the event will have the virus. This is not the same as 50% chance that you will catch the virus. It is up to you to assess that level of risk, decide if you want to go and if you do go, what precautions to take – that’s the purpose of this website, to provide you with the information necessary for you to make those decisions based on the available evidence and respond accordingly – evidence based decisions.
Q: Why does the My-Event page start with Dacorum?
A: The page has to start somewhere and I live in the district of Dacorum (Hemel Hempstead, Berhamsted and others) so I selfishly choose to start it there. I thought it better to display some data initially than to load a page with an empty graph.