Covid-19 Infections: City and University of Cambridge (Autumn 2020)
With the return of the students at the beginning of October, an impact on Covid-19 infection numbers in Cambridge was widely expected. As it turns out, infections among University students had hardly any impact on the non-University community for several weeks, but it looks as if any outbreak at the University is followed by an up-tick in city infections a little later.

The "second wave" of Covid-19 in November was somewhat exacerbated by the presence of the students, but the "third wave" in January definitely was not.

Daily Positive COVID-19 Tests per 100 000 Population (further explanations below)
Your browser does not support graphics

Interpretation: Once students returned, infection numbers within the University were rather high but quickly brought under control. Then, at the beginning of November, roughly two weeks after the University's initial peak, city numbers rose rapidly, possible as a consequence. Shortly afterwards, University numbers went up sharply, but were again very quickly brought under control and are were almost at zero in early December. Assuming that all University staff and students get tested by the University and not elsewhere, the fact that the overall city numbers were falling much more slowly than the University's implies that net numbers in the city were risingi by late November, and keep doing so, if at a slower rate than elsewhere.

What does this chart show?
The University of Cambridge started an ambitious Covid-19 testing programme at the beginning of full term, which allows us to study their data in more detail and analyse their impact on Cambridge as a whole. We have no information about Anglia Ruskin University students, whom we treat as part of the general Cambridge population.
The curves
"Greater Cambridge", "Cambridge", "England", "South Cambridgeshire" and "Cambridge University" are the same curves as in the Regional numbers chart.

"Cambridge minus University" shows the Cambridge data, as the red curve, but with the number of positive results from the two University testing programmes subtracted and the Cambridge population () reduced by the University population () for all dates between 4 Oct and 6 Dec. The same was done to obtain the "Greater Cambridge minus University" curve. This is because positive tests obtained by the University also feed into the government figures and therefore also show in the regional data. We wanted to separate these out.

"Cambridge University" shows positive tests taken under the University's testing programme from people who show Covid-19 symptoms and are either staff or students at the University. This programme started operating on 5 Oct and appears to have ended on 6 Dec.

"Cambridge University asymptomatic" shows positive tests taken under the University's asymptomatic testing programme of college-resident undergraduate and postgraduate students who show no Covid-19 symptoms. This programme also ran from 5 Oct to 6 Dec. The number of students tested is reported explicitly, and it is this number on which we base the "per 100000" quotient. In contrast to other testing, the participants are randomly chosen from each student "household" and represent a very good sample (about one third of the total number of resident students each week). These are the first good data we know of that show the actual prevalence of infections in a good sample of an albeit very specific population.

Data averages and manipulation
The curves for "Greater Cambridge", "Cambridge", "South Cambridgeshire" are the seven-day averaged and otherwise manipulated as described for the regional chart. The same is done for the "... minus University" graphs after the University data have been subtracted.

University data are reported weekly. If no new data have yet been reported, we use the previous week's rate.

Conclusions and caveats
The main conclusions of this study of the University's testing data are