John Burn-Murdoch Profile picture
Stories, stats & scatterplots for @FinancialTimes | Daily updates of the coronavirus trajectory tracker | | #dataviz
AlanLB 🦋 ⚖ 🦋 Profile picture Kevin Coates Profile picture eDo Profile picture Adam Smithee Profile picture Glenys M Goodwill Profile picture 45 added to My Authors
3 Aug
NEW: I’ve updated the chart of new cases in England & Wales to show two critical parts of the story:
• Differentiating between places with an isolated cluster (Swindon’s Iceland depot) vs community spread
• Showing where Leicester was when it locked down… Image
By incorporating that crucial detail of isolated cluster vs community spread, the data and chart better reflect PHE’s own "watchlist" of local areas (see here…)

They also show why Liverpool and Swindon were exempt from the blanket lockdowns. Image
And by adding Leicester at lockdown for context, we can show what @robertcuffe explained so concisely last week:

With the new restrictions in the NW, govt is acting *earlier* than they did in Leicester, but is also acting more *lightly* than it did there
Read 4 tweets
3 Aug
Some people are using the US rowers story to say outdoor activities are not Covid-safe.

The rowers were all infected by their physical therapist, who had prolonged, close, indoor contact with "pretty much the whole team" in his sessions…
This is *totally* different to spending a day on the beach or in the park, even if using shared bathrooms, shower areas, car parks.

Yet these pictures are being used to illustrate stories about the resurgence of the virus. It's hugely misleading and discourages safe behaviour.
A study of 318 outbreak clusters found only one in an outdoor environment…

Another one found that the risk indoors is approximately 19x the risk outdoors…
Read 4 tweets
14 Jul
NEW: today we’ve launched a live dashboard tracking the UK’s economic recovery in real-time.

We’re using fast data (lagged by as little as one day) on transactions, footfall & job ads to show which sectors are bouncing back and which aren’t.

Full piece:
So far, the overall picture is of a slow and abortive recovery.

Total consumer spending was down year-on-year by almost 40% as the lockdown bit. It has recovered half of that loss, but has been stuck at around -20% y-o-y for weeks with no sign of a renewed rebound.
We see a similar picture in retail footfall, where the reopening of non-essential shops on June 15 sparked a sharp uptick, but since then progress has been minimal.

Visits to shops are still -40% year-on-year, well below what we see in peer countries like Sweden & Italy.
Read 10 tweets
13 Jul
NEW: we’ve updated our excess mortality tracker, the best measure for international comparisons of Covid deaths

- US hit 149k a month ago, highest worldwide by far
- Peru & Ecuador now have world’s highest *rates*
- South Africa into excess

Free to read:…
Here are all countries’ excess deaths according to different metrics:
- Population-adjusted rate
- Raw numbers
- % of historical baseline

Peru & Ecuador among the worst on most metrics, UK & Spain also very hard hit. US has passed France for excess death rate, and still rising.
We’ve added new cities.

Whole top tow of hulking red peaks is Latin America.

@jude_webber & I wrote about Mexico City’s soaring death toll, which is hugely under-counted in official figures…

Mexico City & Lima are now 2nd and 3rd behind NYC worldwide.
Read 10 tweets
10 Jul
NEW with @jemimajoanna: why are so few people in the UK wearing masks?

We dug into the numbers, the science and the nuance to see what’s going on

Story here:…

Thread follows:
1) Months into an airborne disease epidemic that has already killed tens of thousands, only 36% of Brits are wearing masks in public spaces.

Far lower than vast majority of peer countries, including all of Asia, most of Europe, & US.

Only Nordics & Aus are even more mask-averse
2) Why?

A big factor is masks are still not mandatory in UK public spaces.

This despite dozens of studies showing that masks — including homemade ones — reduce transmission risk…………
Read 6 tweets
6 Jul
A lot of people are asking how "backward tracing" is any different to standard contact tracing.

It’s not obvious, but there are major differences, and it’s one of the factors cited as being key to Japan’s success at keeping the virus under control.

Here comes a quick thread:
1) Standard contact tracing moves *forward* from the positive case, asking who they have come into contact with *since they got the virus*, seeking to find out who they could have passed it onto, so those people can be isolated.
2) Backward tracing is the other way around: it starts with the positive case and works *backwards*, seeking to find out who could have infected *them*.

This is key with a virus like Covid-19 where most people don’t infect anyone else, but a few people infect *lots* of people.
Read 8 tweets
4 Jul
This had been reported on by the MSM multiple times over several weeks, and it doesn't/didn't affect Leicester at all.

I'm all for legitimate criticism of bad data, but this is incorrect and based on a misunderstanding.
As I explained on Thursday, we knew the consolidation of pillar 1 and pillar 2 data would reduce the total case count. @BBCMoreOrLess (the MSM) first reported this over a month ago
Leicester was locked down based on trends in *that consolidated series of pillar 1+2 data*, where no individual was counted twice.

Claims like the one here just had to the sense of uncertainty and alarm, but are completely false and unnecessary.
Read 4 tweets
2 Jul
Credit where it's due, PHE have now published full pillars 1+2 data for all English local authorities, going back to the start of the outbreak.

Available to download from the main UK government dashboard…
As part of the update process, total UK-wide cases have fallen by around 30,000 and now stand at 283,757.

This was anticipated, as we knew (thanks to @BBCMoreOrLess) that there was a double-counting issue with the testing data, and those duplicates have now gone.
Happily, rather than just subtracting 30k tonight and leaving the old data unchanged, the full time series has been revised, so we now appear to have consistent data on cases by specimen date right back to February (for the UK, England, and English subnational areas).
Read 4 tweets
1 Jul
NEW: one day after we highlighted missing pillar 2 cases, PHE has published full data (both pillars) on new cases in English local areas (as of 10 days ago)

Here are latest cases across whole UK…

By me, @SarahNev, @AndyBounds, @Laura_K_Hughes, @muredickie
The new English data is cases-per-100,000 people during the week ending June 21.

Leicester’s new outbreak makes it the clear hotspot in England, though Merthyr Tydfil is currently recording the most new infections in the wider UK due to the outbreak at its meatpacking plant.
More broadly, some patterns are visible:
• Most areas with high new infections are in Yorkshire or the north & midlands of England
• Wales is counting the cost of 3 meatpacking plant outbreaks
• Northern Ireland and particularly Scotland are seeing very low infection rates
Read 9 tweets
30 Jun
NEW with @AndyBounds, @sarahnev & @Laura_K_Hughes:

The UK government’s published numbers of new cases at local authority level only include pillar 1 and *not* pillar 2 cases, meaning as many as 90% of new cases are missing from the data…

There’s a lot of concern about an uptick in infections in Leicester over the last few days, with the city now going back into lockdown, but here’s what you see if you look at the government’s published data on new cases in Leicester

What new outbreak?
But the govt is sitting on a much more complete dataset including pillar 2, which tells a completely different story

More than 90% of new cases in Leicester are now under pillar 2, yet these numbers are not public. If you’re a local biz owner in Leicester, you only see the blue.
Read 10 tweets
26 Jun
Today we reported on a new @YouGov survey which paints a stark picture of racism in modern Britain.

Far from being an occasional, small or indirect issue, a majority of *all* ethnic minorities in UK have had a racial slur used *to their face* at least once, often multiple times.
The survey also finds racism is highly pervasive in British society, extending into all parts of daily life, including workplaces.

Black Britons are just as likely to have experienced racism in the workplace as in the street (50%).
In any unequal society, the workplace is a key engine of structural discrimination; the data suggest the UK is no different.

A majority of black Britons believe their career development has been held back because of racist attitudes, as do large shares of other minority groups.
Read 4 tweets
23 Jun
NEW: we’ve updated our excess mortality tracker, the best measure for international comparisons of Covid deaths

• US hit 122k a month ago, by far the highest worldwide
• UK 2nd on 66k
• Added Ecuador, world’s highest excess death *rate*

Free to read:… Image
Here are all countries’ excess deaths according to different metrics:
• Population-adjusted rate
• Raw numbers
• % of historical baseline

UK, Spain, Ecuador and Peru among the worst on most metrics. Image
Here’s our data for hardest hit cities & regions worldwide

Top row is all Latin America, which is now the undoubted global epicentre of the virus. Mortality has soared in the major cities of Ecuador, Brazil, Peru and Chile.

Also updated Moscow, where deaths are soaring. Image
Read 12 tweets
23 Jun
The Adria Tour has been a tragic lesson in rushing the return to normal in sport.

Novak Djokovic organised a tennis tournament where adherence to social distancing has been minimal. He and several others, including one player’s pregnant wife, have now tested positive for Covid.
When I say "adherence to social distancing has been minimal", the players went out clubbing in Belgrade last week after the tournament finished.

Obviously wishing everyone a full and fast recovery, but decision-making throughout this one was dreadful
Players testing positive are young and healthy, but a) that doesn’t mean invulnerable and b) given the amount of close contact w/ fans, it’s possible more vulnerable people infected.

Hopefully no lives lost, but this should be a stark lesson to all events businesses regardless.
Read 4 tweets
22 Jun
NEW with @wwftuk:

An exploration of the factors that have led ethnic minorities to be harder hit by Covid than the white population in the UK


Thread of charts and other evidence follows:
1) Most of the numbers we hear on Covid & BAME groups refer to death rates — Covid deaths divided by population — but it’s more useful to think about Covid & minorities in two steps:
• Infection rate (exposure to the virus)
• Case fatality rate (death rate among those infected)
2) There have been many studies on deaths/population rates, but fewer on infection (exposure) rates.

But a recent @QMUL study found that among 1.3m people across four London CCGs, black and South Asian people were much more likely than white people to get infected with Covid.
Read 16 tweets
21 Jun
NEW with @muradahmed:

With Liverpool on the brink of an historic title, we took a deep dive into their transformation under Klopp both on and off the pitch…

Chart thread follows:
1) Liverpool’s progression has been remarkable. Here they are compared to all clubs in Europe’s big leagues over 20+ years.

Since Klopp’s arrival in 15-16, they’ve gone from middle of the pack to elite.
2) This season they’ve been especially efficient, eschewing blowout victories and instead shutting down matches after taking a narrow lead.

Liverpool have 13 more points this season than a team with their goal difference would typically achieve.
Read 12 tweets
15 Jun
2) Proximity of encounter matters, but it’s tangled up with duration, repetition & ventilation.

People worry about a runner passing them too close, but half a second of 1m contact outdoors is less risky than several minutes of 2m contact, especially with poor ventilation.
3) Plenty of studies showing masks reduce risk.

A Lancet meta-analysis found masks reduce infection by 80% (similar to effects of 1m or more of distancing) but NB this was based on surgical grade masks

With homemade or basic cotton masks, effect will be smaller, but still helps
4) Evidence also points to raised voices, singing etc as being a risk factor, as these increase the likelihood of spreading more droplets and spreading them further.

A study found one infected person infected 52 others at choir practice.
Read 15 tweets
10 Jun
It was March 11 when we put out our first Covid-19 trajectory chart, which highlighted:
• UK already on same course as Italy w/ confirmed cases doubling every ~3 days
• UK trailing Italy by more like 2 weeks, not 4 as stated by Vallance the next day
• The data was right here
Here’s the original tweet.

We updated this every day for weeks on end, consistently showing that UK’s doubling-time was faster than decision-makers were saying.

Four days later we added Covid deaths curves, which again unequivocally showed UK was on a steep doubling path and trailing Italy by a much shorter period than we were hearing from officials.

We were putting these charts out for 12 days before lockdown.
Read 4 tweets
10 Jun
NEW: we’ve updated our excess mortality tracker, the best measure for international comparisons of Covid deaths

• UK had 64,000 more deaths than usual to May 29, highest excess death rate globally
• US hit 100k a month ago
• Added Brazil

Free to read:…
Here are all countries’ excess deaths according to different metrics:
• Population-adjusted rate
• Raw numbers
• % of historical baseline
Here’s our data for hardest hit cities & regions worldwide.

Top row is all Latin America, which is now the undoubted global epicentre of the virus. Mortality has soared in the major cities of Ecuador, Brazil, Peru and Chile.

Jakarta also still showing signs of elevated deaths.
Read 10 tweets
4 Jun
NEW: much has been made this week of Spain recording zero new Covid deaths for two successive days.

Spanish Prime Minister Pedro Sanchéz called it "A success for all".

Unfortunately it’s also nonsense.

Story by me & @danieldombey:…

Thread follows:
1/ On Sun, Mon & Tues Spain’s total cumulative number of Covid deaths remained flat on 27,127.

This was widely reported as zero new deaths for two days in a row, and indeed that line was trumpeted by the Prime Minister
2/ But if you look at the Spanish govt’s daily data releases, you’ll spot a new footnote that appeared around a week ago:…

"Only the cases in which the date of death is the day before the time of writing this report are added to the total on a daily basis"
Read 15 tweets
2 Jun
NEW: we’ve updated our excess mortality tracker, the best measure for international comparisons of Covid death tolls

UK had 62,000 more deaths than usual through to May 22, the highest rate of excess deaths in the world

Free to read:…
Here are the full figures for all countries’ outbreaks.

Across much of Europe, mortality is now back at or approaching historical baselines, but in the US deaths are still significantly elevated, and in Latin America excess deaths continue to rise steeply.
Indeed, Peru has become one the new global epicentres, with officials struggling to capture its full death toll.

Nationally, Peru’s mortality is 2x usual levels, and the capital Lima is among the hardest hit subnational areas worldwide. Several Brazilian cities are also reeling.
Read 10 tweets
1 Jun
NEW: we’re pleased to announce all of our excess deaths data is now public, in CSV format…

• Data for 24 countries and counting
• Down to regions/cities in many cases
• Includes numbers for Italy through to end of April
We’re not the only ones collecting and publishing this data.

Worth checking out the resources already published by @J_CD_T & co at @TheEconomist (…) and @atmccann & co at @nytimes (…)
We’ve already shared this data directly with academics, but we’re taking the step to make it completely public today so that more people can dig into the numbers, combine with their own datasets etc.
Read 6 tweets