Corona Daily 051: On the Recovery Road

Throughout the pandemic, I have been running a private research, just for myself. I ask the self-employed, my businessmen friends, restaurant owners, carpenters, taxi drivers, shopkeepers what their recovery has been like. Usually they give a percent figure. Last year, when lockdowns were strict, the percent never went beyond 30%. Now the luckier ones talk about 50%. Only a single friend, a sales manager for a medical equipment company said their sales have grown 250%. He has been working day and night, trying to meet the demand.

*****

Yesterday, YELP has published in a formal manner similar data for the USA. Yelp is a San Francisco company that is a data king. It publishes crowd-sourced reviews about restaurants and other businesses. There are more than 200 million reviews on its business listing pages. Recovery of consumer interest can be measured by counting the actions consumers take to connect with businesses on Yelp: the number of time a page is viewed, photos or reviews posted. If reviews for a certain restaurant suddenly double, it is safe to assume its business is recovering well. In this research Yelp analysed millions of data points across 1200 categories in all 50 American states.

*****

One big danger of the growth numbers you see in 2021 is that they are compared to the lockdown period of the previous year. The April-June quarter in particular was so abnormally depressed last year; that year-on-year numbers in April-June 2021 for anything will look rosy to miraculous. That growth number can be fallacious.

To avoid that pitfall, Yelp studied consumer actions in three periods: (1) Pre-pandemic: March-May 2019 (2) The onset of the pandemic: March through May 2020 and (3) the start of recovery: March through May 2021. The 2019 pre-pandemic period is used as a baseline, with the 2019 data expressed as 100%. Below, in brackets are the 2021 percent as compared to 2019 (and not 2020).

*****

After a dramatic drop in 2020, American restaurants have now bounced back to 86% of 2019. Food business recovered to 92%. People are less worried about in-person dining as well as shared food experiences. Conveyer belt sushi (97%), tapas (82%), hot pot (75%), dim sum (84%) have rebounded. Buffets, not surprisingly, have recovered only 52%.

While this is happening, people can’t give up the lockdown addiction of home orders. Food delivery services are 189% compared to 2019.

*****

Americans are out for revenge shopping. Personal savings level is at their highest since the 1960s. Consumer interest in shopping is almost complete, it is 92%.

Shoes (79%), jewelry (88%) and clothing (76%) are doing well. As far as shopping venues are concerned, antiques (100%), flea markets (94%), vintage shops (87%) and charity shops (101%) have reached the consumer interest levels of 2019.

Home service professionals are in high demand, partly because of workspace in houses that needs maintenance or upgrading. Consumer interest in carpenters (151%), handyworkers (147%), plasterboard (147%), refinishing (133%), and tiling (132%) is booming, surpassing pre-pandemic levels. The heated housing market in the USA has completed recovery for real estate agents (104%) and mortgage brokers.

Many Americans are presumably planning to return to offices. Shoe repair (77%) and watch repair (99%) services suddenly surged in the past three months. While childcare services (71%) are slowly recovering, the demand for pandemic pet care has grown as shown by interest in pet sitting (114%), and pet boarding (121%).

With people itching to leave houses, interest in campervan dealers (125%), motorcycle dealers (126%), and used car dealers (110%) presents a picture of America on the road.

*****

In terms of geography, Idaho, Wyoming and South Dakota (110%) are the leaders in recovery. Florida (98%), Georgia (91%) and Texas (91%) are getting there.

Washington DC (54%), New York (63%) and Massachusetts (71%) are the laggards. They have densely populated metro areas, and lingering pandemic restrictions.

I must point out that the Yelp report data is about consumer interest, not necessarily consumer action. That is measured by actual purchases of goods and services. However, consumer interest is a good indication of the direction of the recovery graph. It shows that the American consumer sentiment is optimistic again.

Ravi 

Corona Daily 052: Where Do You Get Your News From?

Where do we learn about the crime stories from our city, the latest lockdown restrictions, local and global politics? My parents, in their eighties, watch Indian TV channels and read two mainstream local newspapers. My wife reads an English language newspaper, and uses WhatsApp. I digitally subscribe to a few newspapers, but read them in a non-linear fashion. My 17-year old daughter gets news on her mobile before anybody else in our family does.

*****

Based on data from six continents and 46 countries, the Reuters Institute Digital News Report 2021 was published yesterday. It tells us how the coronavirus pandemic has quickened earlier trends. The pandemic shook the news industry, particularly printed newspapers. Initially the lockdowns severely restricted their distribution. More worrying was their loss of advertising revenue. There have been large scale layoffs in the news industry. In the USA, sixty local news organizations were shut off. Free sheets like Metro or the Standard in London have lost 40% of their income. New business models ask readers to subscribe, take membership or donate. These attempts have been unsuccessful in replacing the lost ad revenue.

*****

 The good news for traditional media is that the trust gap between news media and social media has grown. TV channels like BBC news inspire trust in their viewers. Trust levels vary. Finland has the highest overall trust in the news (65%), and USA has the lowest (29%). After 2016, the US news media experienced a Trump bump, now it is replaced by a Trump slump. Both Fox News and CNN have lost large number of viewers. (Biden is boring).

*****

Young people and people with low level of education use social media as their source of news. TikTok now reaches 24% of under-35s, many of them using the platform for news.

Generation Z (18-24) has little connection to newspapers, websites or news apps. They get their news from social media, aggregators (Google news, Apple news) or mobile alerts. Mobile aggregators play little role in the Western countries, but are strong in India, Indonesia, South Korea and Thailand. AI-powered apps like Daily Hunt, Smart News, Naver, and Line Today are important sources of discovering news.

Most young people across the world feel they are not fairly represented in mainstream media. There is little coverage of the issues they care about. That is why they embrace social media. They neither pay money nor attention to the media that is unfair and uninteresting for them.

Online newspaper subscribers tend to be more educated, richer and older. As a rule, nobody under 40 digitally subscribes to a newspaper.

*****

Overall, the leading channel of misinformation is Facebook (28%), followed by news websites (17%), WhatsApp (15%), Twitter (6%) and YouTube (6%). In India, Brazil, Indonesia, South Africa and Nigeria, WhatsApp is the big culprit. Its closed, encrypted nature prevents fact-checkers from spotting the origin. UK and USA cite Facebook as the prime worry for misinformation.

*****

Smartphone dependence has grown during lockdowns. Now 73% access news via smartphones. Mobile phones are rapidly replacing laptops and desktops. In the pandemic, government has further boosted smartphone use for covid messages and vaccine appointments. In the UK, the computer-smartphone gap has grown to 25 points. In the past ten years, computer use fell from 67% to 43%. Smartphone use went up from 29% to 68% in the same period.

*****

When asked about their motivation, people said Twitter was good to access and debate news. Facebook was used for other reasons. Instagram, Snapchat and TikTok are seen as fun and time-killers. YouTube users were looking for alternative perspective on the news.

*****

Aggregators like Google news or Apple news simply take you to links. Some journalists still have to create original content. If nobody under forty is interested in reading news in the mainstream media, printed newspapers have another fifty years to live.

Ravi 

Corona Daily 053: Unhappy Birthday to You

Since last March, there is another epidemic taking place – that of academic studies and research papers. Since we live in emergency times, every researcher is keen to come up with creative ideas, test them quickly, and get a paper published. Normally, reputed science journals insist on greater scrutiny and peer reviews. Not any more. Just like drugs and vaccines, research papers are granted emergency authorizations.

Many studies, though of little practical relevance, are highly amusing. A paper published on 17 September 2020 described the effect of singing and talking on the spread of covid-19. Twelve volunteer singers, including seven professional opera singers (2 basses, 1 baritone, 2 altos, and 2 sopranos) were invited. I presume the researchers as well as the opera singers were out of work. The researchers, who were from Sweden’s Lund university, observed the trajectory of aerosols emitted by the twelve singers singing different songs. During the tests, singers sang into a funnel. The aerosol particles were measured at the other end of the funnel.

At the beginning of every epidemic, scientists ask people to wash their hands thoroughly. To time the hand washing with soap, you are expected to sing the ‘happy birthday’ song in full.

The September study discovered that the happy birthday song can cause wider transmission, because the lyrics are heavy on consonants, and the hard Ps and Bs send more droplets into the air.

*****

 Compared to that entertaining study, the paper published this week is serious and clever. In an intelligent way, it investigates what might have happened in three million American households without entering a single house.

The researchers gathered private health insurance data from nearly 3 million households in the USA. The data included the birthdates of everyone in the house. This was the first data point.

They also collected the administrative health care data giving the covid-19 positive test records. The time frame was 1 January to 8 November 2020. This was another data point.

This large scale study was approved by authorities for waiver of informed consent. The data was de-identified, meaning researchers knew only code numbers and not the names of the individuals.

The two data points, birthdays and the covid-19 positive tests were studied together to see if there was any correlation between them.

*****

The data analysis revealed that in counties with high rate of transmission, households were 30% more likely to test covid positive within two weeks after someone had a birthday. In households where a child’s birthday occurred, the incidence was even higher. Out of 10,000 people, almost 6 excessive positive tests were found within two weeks of an adult’s birthday, and nearly 16 excessive positive cases after a child’s birthday.

In formal settings, such as corporate or political meetings, people take more precautions, with masks and distance. However, when we are with our close family or friends in an informal setting, our guard is down. It is harder to imagine our infecting our own family members.

The world has been focused on mass gatherings, crowds and superspreader events. This study showed that even under strict lockdown conditions, a large number of American families was secretly celebrating birthdays, particularly kids’ birthdays. After the child blew out the candles spreading aerosols on the large cake, the family and friends sang the happy birthday song with its hard Ps and Bs, completely contaminating the room.

***** 

From a researcher’s viewpoint the beauty of the birthdate is that everyone has one. In places where infections are still raging, people will do well to read this research paper. That should hopefully stop them from arranging a secret birthday party during a red-level lockdown.

Ravi 

Corona Daily 054: How will the Covid Pandemic End? When? Part Final

Today, I come to the most exciting part of the pandemic forecasting exercise – predicting the precise date when the pandemic ends. By ending I mean social ending, with no masks, no social distancing, crowded bazaars and trains, and traffic jams. There will still be a few cases and deaths, but they will not be front page news. Remember, 20,000 people died of respiratory diseases (flu and pneumonia) EVERY DAY in 2019. The official daily covid-19 figure has never surpassed that count.

*****

WHO says the pandemic won’t end unless 70% of the global population is vaccinated. Considering that 30% of global population (younger than 16) doesn’t yet have a vaccine developed for them, this is a gloomy forecast. WHO is probably thinking of eradicating the disease.

Dr Fauci predicts that 75-80% vaccination would allow America enough herd immunity to exit the pandemic before 2021 ends.

Renowned Indian astrologer Bejan Daruwalla projected the pandemic’s end in May 2021. But one year before verifying it, he died of suspected covid in May 2020.

***** 

From my reading, I can recommend two sources that have done serious work on the subject.

One is an article by McKinsey & company dated 26 March 2021. Based on seven variables, the company predicts the pandemic will end in the UK and USA in the second quarter of this year, EU in the third quarter and the rest of the world in 2022.

The other source is an academic paper by a group of computer engineers from Saudi Arabia, Egypt and Ethiopia. Their “covid-19 pandemic prediction study” is based on machine learning models. The model uses ARIMA and SARIMA, which are not the names of Egyptian queens, but acronyms. ARIMA is AutoRegressive Integrated Moving Average. S in SARIMA is Seasonal.

For some odd reason, if you love reading papers with tables and formulas, you may want to read their two papers, one published in October 2020 and another in April 2021. Their projected pandemic end dates are November 2021 for USA, India, Europe, UK, and September 2021 for Brazil. Unfortunately, after their April paper, severe waves happened in India and Brazil. They made the predictions unrealistic, certainly for those two countries. The Saudi-Egyptian model ignores vaccination levels.  

*****

My method is a simplified one. I have assumed a bell curve (remember normal distribution?) where the virus after peaking, declines in a symmetrical fashion so that the peak is exactly in the centre of the bell curve. (Since all models can go wrong, simple is better than complex).

I give here the dates of peak cases and peak deaths for nine countries. The last column is the latest vaccination data until yesterday. First figure is the % receiving at least one shot, second % fully vaccinated. (If you are number-immune, you may jump to the next table.)

Source: Worldometer	Peak cases (date)	Peak deaths (date)	Vaccination % (one shot/fully)
World	902,022(23 Apr 21)	17,498(20 Jan 21)	Negligible
USA	305,067(8 Jan 21)	4,463 (12 Jan 21)	53/45
Brazil	98,135 (18 Jun 21)	4,211 (6 Apr 21)	30/11
India	414,433 (6 May 21)	4,930 (15 May 21)	16/3.6
Russia	29,935 (24 Dec 20)	635 (24 Dec 20)	14/11
UK	67,808 (8 Jan 21)	1,823 (20 Jan 21)	65/47
France	58,581 (6 Nov 20)	932 (13 Nov 20)	48/25
Germany	32,546 (14 Apr 21)	1,244 (29 Dec 20)	51/31
Poland	35,246 (1 Apr 21)	954 (8 Apr 21)	43/29

 

(Note: Brazil has reached another peak this week. We don’t yet know when Brazil will peak. Russia’s figures are not trustworthy. Both cases and deaths peaked in December, but the curve was suspiciously flat after that and is now growing. Russians don’t trust vaccines either. Only 14% have opted to take a shot despite plenty of supply. Germany is an anomaly in that the deaths peaked in December, but cases had another surge in April.)

*****

Taking the peaks from the table above as the mid-point between the first case and the pandemic ending, I get the following pandemic end-dates.

	First Case	Peak	End date
World	1 Dec 2019	23 April 2021	16 October 2022
USA	22 Jan 2020	8 Jan 2021	25 Dec 2021
Brazil	26 Feb 2020	18 June 2021	10 October 2022
India	30 Jan 2020	6 May 2021	12 August 2022
Russia	31 Jan 2020	24 Dec 2020	17 Nov 2021
UK	31 Jan 2020	8 Jan 2021	16 Dec 2021
France	24 Jan 2020	6 Nov 2020	19 Aug 2021
Germany	27 Jan 2020	14 April 2021	1 July 2022
Poland	4 March 2020	1 April 2021	28 April 2022

 

There you have the joyful dates: Christmas 2021 for the UK and USA. 12 August 2022 for India. I am more confident about dates for these three countries than others. There may be additional waves. As long as they don’t alter the peak; the end date should not change much. This assumes the weakening of the virus severity because of the ‘self-limiting’ nature, natural immunity and vaccination.

If your country is missing from the table here, find the date of the first case (Wikipedia), the peak (worldometer), and calculate the end date taking the peak as the mid-point. The result would be more scientific than an astrologer’s prediction.

Ravi 

Corona Daily 055: How will the Covid Pandemic End? When? Part Two

One way to predict the end of the current pandemic is to analyse how the earlier pandemics ended.

Infectious diseases have threatened humans since the Neolithic revolution 12,000 years ago, when human hunter-gatherers settled in villages to domesticate animals and cultivate crops. Animal domestication was man’s earliest systematic and large scale manipulation of nature. The man-animal interaction was the cause of several pandemics. Justinian plague (541-549 AD) wiped out half the world’s population, Black Death (1346-1353) killed over 50 million. Not enough reliable documentation exists to know how they ended. Probably everyone who was not dead had developed immunity through natural infection.

The current pandemic raised awareness about the Spanish influenza of 1918. Without vaccines and medicines how did the virus disappear? Does Wikipedia offer the end-date of that pandemic the way it mentions the dates of individual deaths?

Readers may be shocked to learn the Spanish flu virus never disappeared. It caused minor pandemics in 1957, in 1968 and in 2009. The H1N1 virus gave rise to swine flu in 2009 killing 284,000. During the Spanish influenza, the same virus was far deadlier killing 50 million people.

The AIDS pandemic continues for forty years, and has killed over 37 million. We are just less afraid of HIV because it doesn’t spread through sneezing and coughing. Chikungunya (2014), Zika (2015), Ebola in Africa (1976-2021) are all around us and can cause outbreaks at opportune times.

If some of these viruses are not widely experienced, take the case of the more familiar malaria. Transmitted via parasite, it is as old as the human race. In 2018, there were 228 million malaria cases and 405,000 deaths worldwide. But malaria didn’t ever shut schools or restaurants.

*****

No doubt vaccinations help reduce hospitalizations and the death toll, but not a single epidemiologist would think vaccines will eradicate the covid-19 virus.

In human history, the only disease eradicated through vaccination was smallpox. The last person to naturally contract it was Ali Maow Maalin, a hospital cook in Somalia. (In 1977, he recovered from small pox, but eventually died of Malaria in 2013). Smallpox was exceptional in many respects. Its vaccine offered lifelong protection. More importantly, its virus, Variola major, had no animal host. Eliminating it in humans really meant its end. Smallpox symptoms were so unusual, it was easier to trace and isolate patients.

*****  

That is also probably the reason why SARS-I (2002) ended so quickly, claiming only 813 lives. The symptoms appeared together with infection. The infected person could immediately isolate. SARS-2, better known as Covid-19, has a gap of a couple of days between infection and symptoms. Worse, it is possible to attract and spread infection without having any symptoms. The strategies that worked with SARS can’t work with covid-19.

*****

In short, the end of the pandemic can’t be associated with the disappearance of the covid-19 virus. It will stay with us for many decades, perhaps for ever, just like influenza. Of course, it is expected to turn into a seasonal flu virus, emerging every year in a mild form probably in winter months. Winter because children are at school, and everybody spends more time indoors.

It will be a sort of truce agreement between humans and the corona virus.

*****

Historians talk of two types of pandemic endings, the medical ending (like for smallpox) and the social ending, when fear recedes, panic disappears, life returns to normal. When people ask: ‘when will the pandemic end?’, they are asking about the social ending, says Dr Greene, a historian at Johns Hopkins.

Tomorrow, I will share the latest thinking about when this social ending may happen. It means in-person classes at schools and universities, opening of businesses, offices, restaurants, theatres, malls, large gatherings and inter-regional and international travel as it was pre-pandemic.

(Continued tomorrow)

Ravi