israeli_data <- read.csv("https://github.com/addsding/data607/raw/main/extra/ec3/israeli_vaccination_data.csv")
head(israeli_data)This data describes August 2021 data for Israeli hospitalization (“Severe Cases”) rates for people under 50 (assume “50 and under”) and over 50, for both un-vaccinated and fully vaccinated populations. The columns are:
As the assignment is about Israel, here are some high-level facts I’ve discovered to start:
With this as a base, let’s begin by tidying the data for analysis!
When looking at this data, it’s honestly a little hard to see how we could get this one .csv into one table. I honestly think it’d be cleaner to have these be separate tables:
We’ll begin with the first one by pivoting each of the counts by age bucket.
# pivot the counts
israeli_data_pop_count <-
pivot_longer(israeli_data[1:5]
, cols=c(2, 4)
, names_to="population_type"
, values_to="count"
)
# pivot the percentages
israeli_data_pop_count <-
pivot_longer(israeli_data_pop_count
, cols=2:3
, names_to="population_type_"
, values_to="percentage_of_population"
)
# clean the names to remove the periods
israeli_data_pop_count$population_type <-
gsub("\\.", " ", israeli_data_pop_count$population_type)
# this one needs white space removal as there were periods at the end
israeli_data_pop_count$population_type_ <-
trimws(
gsub("\\.", " ", israeli_data_pop_count$population_type_)
)
# filter down the rows
israeli_data_pop_count <-
israeli_data_pop_count |> filter(population_type == population_type_)
# remove the duplicate row
israeli_data_pop_count <-
subset(israeli_data_pop_count, select=-c(population_type_))
head(israeli_data_pop_count)Next is the data frame for counts per 100k.
# pivot the counts
israeli_data_100k <-
pivot_longer(israeli_data[c(1, 6, 7)]
, cols=2:3
, names_to="population_type"
, values_to="count"
)
# clean the names to remove the periods
israeli_data_100k$population_type <- gsub("\\.", " ", israeli_data_100k$population_type)
head(israeli_data_100k)Both data frames look clean and tidy!
We have numbers and what percentage they represent, but I don’t believe that’d be total population for Israel. Let’s double check that though.
For the < 50 group, let’s see what the total number
would be based on both population types.
print(paste("Population based on not vax population:"
, toString(
israeli_data_pop_count$count[1] /
israeli_data_pop_count$percentage_of_population[1])
)
)## [1] "Population based on not vax population: 4793278.96995708"print(paste("Population based on vax population:"
, toString(
israeli_data_pop_count$count[2] /
israeli_data_pop_count$percentage_of_population[2])
)
)## [1] "Population based on vax population: 4796052.05479452"These numbers are very close – we can assume that the difference is
due to rounding, this is relatively accurate otherwise. From this, I’ll
say that the total population of the < 50 group is
4.79m
Now, let’s do the same for the > 50 group.
print(paste("Population based on not vax population:"
, toString(
israeli_data_pop_count$count[3] /
israeli_data_pop_count$percentage_of_population[3])
)
)## [1] "Population based on not vax population: 2355417.72151899"print(paste("Population based on vax population:"
, toString(
israeli_data_pop_count$count[4] /
israeli_data_pop_count$percentage_of_population[4])
)
)## [1] "Population based on vax population: 2360084.07079646"Again, due to rounding, these are slightly off but otherwise close. I
’ll say that for the > 50 group, the population is
2.36m.
Together, this is 7.15m. This, however, does not come close to the 9.451m that I researched beforehand which is supposed to represent the population of Israel. This discrepancy likely comes down to those who are eligible or not and so to rephrase what we are able to find: we are able to calculate the total population of people who are eligible for the vaccine.
I will also say that if you add up all of the population counts, it does not equal 7.15m. The percentages also don’t add up to 100% – I believe a group is missing here. One I can think of would be folks who are partially vaccinated, so for example they only have 1 of the 2 doses.
Given the formula:
1 - (% fully vaxed severe cases per 100k / % not vaxed severe cases per 100k),
we can calculate it like so:
below_efficacy <- 1 - ((israeli_data_100k$count[2] / 100000) / (israeli_data_100k$count[1] / 100000))
above_efficacy <- 1 - ((israeli_data_100k$count[4] / 100000) / (israeli_data_100k$count[3] / 100000))
print(paste("Efficacy for < 50 population:", toString(below_efficacy)))## [1] "Efficacy for < 50 population: 0.744186046511628"print(paste("Efficacy for > 50 population:", toString(above_efficacy)))## [1] "Efficacy for > 50 population: -0.695906432748538"Based on these results, we can see that for folks fifty and under, being fully vaccinated proves to be 75% effective against severe cases. For folks above the age of fifty, the vaccine actually failed to be effective for these cases as its efficacy was negative. Folks over the age of fifty though are likely to have more health problems and be more susceptible to disease so it makes sense that the vaccine wasn’t as effective.
I would say that past a certain age, it becomes much more individualized as to whether a vaccine would help someone or not due to illnesses and other health concerns that come with older age. I think comparisons would work best for people below fifty as it’s easier to generalize that population than the older one.
The data was pretty hard for me to picture as one data frame so I
ended up splitting it into two, however I’m guessing there’s a way to
combine them – the data just wouldn’t be as tidy in my opinion. The
analysis was a little confusing as I wasn’t entirely sure a negative
efficacy percentage made sense, however I see now that it does since the
fully vaccinated severe case count in the > 50 group is
higher than the non vaccinated one which means that the vaccine wasn’t
effective for this age group.
Next steps would perhaps be to take a look at other countries with similar size / conditions to see if this trend continues here or if Israel is an outlier with its performance.