Vaccine | Tengku Hanis

Some COVID-19 plots for Southeast Asian countries

Wed, 10 Nov 2021 00:00:00 +0000

Recently, I found a GitHub repo containing a global COVID-19 dataset. I thought, why not try to do some plotting for Southeast Asian countries. So, I downloaded the data and limited the data to Southeast Asian countries only (Brunei, Indonesia, Malaysia, Philippines, Singapore, Thailand and Vietnam). I have uploaded this restricted data to my GitHub repo.

We are not going to do anything fancy, just some visualisations.

Let’s begin by reading the data.

library(tidyverse)
covid_sea <- read_csv("https://raw.githubusercontent.com/tengku-hanis/data-owid-covid/main/covid_sea.csv")

We are going to compare between each Southeast Asian countries in terms of:

Daily cases
Daily deaths
Daily tests
Daily vaccinations

Before that, we need to make a function, as all the above items have a generic things to plot with the exception on y axis.

easy_plot <- function(var1, lab_title, yaxis_lab, span = 0.14){
  covid_sea %>% 
    select(date, location, {{var1}}) %>% 
    drop_na() %>% 
    ggplot(aes(date, {{var1}}, color = location)) +
    geom_smooth(se = F, span = 0.14) +
    geom_point(aes(color = location), alpha = 0.2) +
    geom_line(aes(color = location), alpha = 0.2, linetype = "dashed") +
    labs(title = {{lab_title}}) +
    ylab({{yaxis_lab}}) +
    xlab("Date") +
    theme_minimal() 
}

var1 is going to be the item/variable that we want to compare, lab_title is the plot title, yaxis_lab is the label on the y axis, and span is just how smooth our smoothen line should be.

Daily cases

easy_plot(new_cases, "Daily cases for southeast Asian countries", "Daily cases", span = 0.8)

We cannot compare in terms of the frequency as big countries like Indonesia is expected to had a higher number of daily cases. A smoothen line though very basic, may indicate a simple trend. Thailand, Malaysia, Philippines and Indonesia seems to had a decreasing trend of cases. On the other hand, the daily cases in Vietnam seems to start to increase. Singapore had a more stabilised trend of cases, though a higher number of cases was observed in the latest period. Lastly, Brunei had too little cases, for us to see any sort of trend at the scale of the between countries comparison.

Daily deaths

easy_plot(new_deaths, "Daily deaths for southeast Asian countries", "Daily deaths", span = 0.8)

Philippines and Indonesia seems started to had a bit of increasing trend. Other countries look okay.

Daily tests

easy_plot(new_tests, "Daily tests for southeast Asian countries", "Daily tests", span = 0.2)

The daily tests plot looks a bit weird for Vietnam. Actually, the daily tests below zero are not avaliable (not sure if there is no test done in the period or the values is just missing). Hence, the weird looking plot for Vietnam. Data for Brunei and Thailand are not available. Malaysia seems to be quite aggressive in COVID-19 testing, even on par with Indonesia. Also, Vietnam seems to be very aggressive in the latest period, probably to cover the lack of COVID-19 testing previously.

Daily vaccinations

easy_plot(new_vaccinations, "Daily vaccinations for southeast Asian countries", "Daily vaccinations", span = 0.9)

Malaysia and Singapore had quite a similar distribution. Vietnam, Philippines, Thailand and Indonesia quite similar in which they had a series of wave in the rate of vaccinations, though the trend of wave for Thailand is less obvious. Again, the number in Brunei was too little for us to see any trend or distribution at this scale.

Malaysia situation

Let’s do a plot, specific to Malaysia. We going to scale the numbers, so that we able to see a comparison in term of trend or distribution.

covid_sea %>% 
  filter(location == "Malaysia") %>% 
  mutate(new_cases = scale(new_cases), 
         new_deaths = scale(new_deaths), 
         new_tests = scale(new_tests), 
         new_vaccinations = scale(new_vaccinations)) %>% 
  ggplot(aes(date)) +
  geom_line(aes(y = new_cases, color = "new_cases"), alpha = 0.3) +
  geom_line(aes(y = new_deaths, color = "new_deaths"), alpha = 0.3) +
  geom_line(aes(y = new_tests, color = "new_tests"), alpha = 0.3) +
  geom_line(aes(y = new_vaccinations, color = "new_vaccinations"), alpha = 0.3) +
  geom_point(aes(y = new_cases, color = "new_cases"), alpha = 0.3) +
  geom_point(aes(y = new_deaths, color = "new_deaths"), alpha = 0.3) +
  geom_point(aes(y = new_tests, color = "new_tests"), alpha = 0.3) +
  geom_point(aes(y = new_vaccinations, color = "new_vaccinations"), alpha = 0.3) +
  geom_smooth(aes(y = new_cases, color = "new_cases"), se = F, span = 0.3) +
  geom_smooth(aes(y = new_deaths, color = "new_deaths"), se = F, span = 0.3) +
  geom_smooth(aes(y = new_tests, color = "new_tests"), se = F, span = 0.3) +
  geom_smooth(aes(y = new_vaccinations, color = "new_vaccinations"), se = F, span = 0.6) +
  labs(title = "Situation in Malaysia") +
  ylab("Scaled Frequency") +
  xlab("Date") +
  guides(color = guide_legend("Items")) +
  scale_color_discrete(labels = c("Daily cases", "Daily deaths", "Daily tests", "Daily vaccinations")) +
  theme_minimal()

Interestingly, as the number of vaccination increased up to a certain threshold, the number of daily cases and daily deaths started to decreased. Obviously, the daily testing also decreased as in Malaysia, COVID-19 testing is done based on suspected cases and their persons of contact instead of mass testing.

Disclaimer: Please take anything written here with a massive grain of salt.

Data source: https://github.com/owid/covid-19-data/tree/master/public/data

COVID-19 vaccine interest in Malaysia

Sun, 17 Oct 2021 00:00:00 +0000

We are going to do a basic google trends search using gtrendsR package and do some plotting with ggplot2.

These are the required packages.

library(gtrendsR)
library(tidyverse)

Run gtrends() function to search our keywords of interest (i.e; type of vaccine). So far, we only used 4 type of vaccines in Malaysia.

vaccine <- gtrends(c("pfizer", "astrazeneca", "sinovac", "cansino"), geo = "MY")

Then, plot our keywords.

plot(vaccine)

Probably, it’s better if we filter our date to when the COVID-19 pandemic started, which is around March 2020.

vaccine$interest_over_time %>% 
  group_by(keyword) %>% 
  filter(hits != "<1" & date > as.Date("2020-03-01")) %>% 
  mutate(hits = as.numeric(hits), 
         date = as.Date(date)) %>% 
  ggplot() + 
  geom_line(aes(x = date, y = hits, color = keyword), size = 0.8) +
  theme_minimal() +
  labs(title = "COVID-19 vaccine interest in Malaysia", y = "Search hits", x = "Date") +
  scale_x_date(date_breaks = "4 month")

So, AstraZeneca vaccine is of high interest, probably due to infamous blood clotting issue. Next, we can also get the search keywords based on the states.

vaccine$interest_by_region %>% 
  group_by(location) %>% 
  ggplot(aes(location, hits, fill = keyword)) +
  geom_col(alpha = 0.8) +
  coord_flip() +
  theme_minimal() +
  scale_fill_viridis_d() +
  labs(title = "COVID-19 vaccine interest in Malaysia by states", y = "Search hits", x = "")

Lastly, we can plot the search keywords based on the city.

vaccine$interest_by_city %>% 
  group_by(location) %>% 
  drop_na() %>% 
  ggplot(aes(location, hits, fill = keyword)) +
  geom_col(alpha = 0.8) +
  coord_flip() +
  theme_minimal() +
  scale_fill_viridis_d() +
  labs(title = "COVID-19 vaccine interest in Malaysia by cities", y = "Search hits", x = "")

gtrendsR with just a bit of plots certainly very useful if we want to gauge certain issues in the community.

How many Malaysian should be vaccinated to get herd immunity from COVID-19?

Mon, 07 Dec 2020 00:00:00 +0000

Recently I have read an article that the Malaysian government have made a deal with Pfizer for 6.4 million Malaysian to be vaccinated. So, I am wondering what is the minimal number of people should be vaccinated.

I have also come across this interesting article, which explains how we can calculate a minimal number of people to be vaccinated to achieves herd immunity based on R naught (R₀).

R naught (R₀)

The basic idea of R₀ or basic reproduction number is quite simple. It describes how many secondary infections will derive from the first case. I think Figure 1 below describes this idea very well.

Figure 1: Basic idea of R₀(image from https://www.atrainceu.com/content/3-basic-reproduction-number-r-naught)

So, R₀ can be affected by a few factors, such as:

proportion of susceptible people at the initial outbreak
infectiousness of the virus or the disease
rate of recovery or death
and a few other factors

As R₀ increases more than 1, the spread of the disease will increases, while R₀ below 1 indicates the spread of the disease will decrease and eventually dies out.

However, I noticed that quite a few including KKM (Ministry of Health, Malaysia) have used the term R₀ in their reports instead of R_e or R_t which is the effective reproduction number or time-varying reproduction number. R₀ refers to the initial reproduction number at the beginning of the outbreak. The “naught” or “zero” in R naught (R₀) is referring to population condition that has zero immunity to the disease.

Herd immunity

Herd immunity is said to occur when a significant proportion of the population is immunized. Subsequently, those whose susceptible (not immunized) will be protected.

How many should be vaccinated

So, back to the initial topic. We can use the formula below to answer this question.

\[P_i > 1 - \frac{1}{R_0}\]

P_i refers to the number of proportion that should be immunized or in this case, vaccinated.

So, after googling, I found one calculation by my lecturer in Biostat Unit, USM, Dr Wan Arifin and his colleague. The R₀ based on his calculation is 2.673. Also, I found another article reported that the R₀ is 3.55 in March, according to KKM.

Malaysian’s population is estimated at 32.7 million by the Department of Statistics, Malaysia (DOSM). So, using the formula above, about 63% to 72% of Malaysian population should vaccinated, and this translates to about 20.6 to 23.5 million people.

The deal that the Malaysian government made with Pfizer is far from enough, but of course, this is a very good and quick decision. We also have other vaccines like Moderna’s vaccine coming up.

Disclaimer: This is just my opinion. Please take it with a massive grain of salt.