Fourth session

compare gapminder wide vs long

gap_wide <- read_csv(“https://bit.ly/gapminder-rsu”))

Load the packages

pacman::p_load(tidyverse, # several datascience packages
               palmerpenguins, # data
               gapminder, #data,
               gtsummary) # for tables)

gap_wide <- read_csv("https://bit.ly/gapminder-rsu")

## Rows: 142 Columns: 38

## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (2): continent, country
## dbl (36): gdpPercap_1952, gdpPercap_1957, gdpPercap_1962, gdpPercap_1967, gd...

## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Data basics

Filtering

Filter by species == Adelie

penguins %>% 
  filter(species == "Adelie")

Filter by two conditions Adelie & island == Dream

Filter body_mass_g < 3000

penguins %>% 
  filter(species == "Adelie" & island != "Torgersen")

You can use other operators beyond just the == operator that tests for equality:

> corresponds to “greater than”
< corresponds to “less than”
>= corresponds to “greater than or equal to”
<= corresponds to “less than or equal to”
!= corresponds to “not equal to.” The ! is used in many programming languages to indicate “not.”

& is “and”, | is “or”, and ! is “not”.

TASK: Adelie but not from Torgersen

TASK Adelie or Gentto

penguins %>% 
  filter(species == "Adelie" | species == "Gentoo")

Another way, using %in%

penguins %>% 
  filter(species %in% c("Adelie", "Gentoo"))

TASK

Filter all penguins from Torgersen weighing more than 3500g

Mutate

Create new variables

bill_ratio = bill_length_mm / bill_depth_mm

penguins %>% 
  mutate(bill_ratio = bill_length_mm / bill_depth_mm)

Relocate after island column

penguins %>% 
  mutate(bill_ratio = bill_length_mm / bill_depth_mm) %>% 
  relocate(bill_ratio, .after = island)

TASK

penguins %>% 
  mutate(body_mass_k = log10(body_mass_g))

Grouping and summarizing data

calculate the mean body_mass_g by specie

penguins %>% 
  group_by(species) %>% 
  # now I will remove the NAs values
  drop_na() %>% 
  # and now the summary of the variable
  summarise(n = n(), 
            mean = mean(body_mass_g),
            sd = sd(body_mass_g), 
            max = max(body_mass_g), 
            min = min(body_mass_g)) %>% 
  mutate(across(where(is.numeric), ~ round(., 1)))

Dealing with NAs

na.rm = T

drop_na()

penguins %>% 
  group_by(species) %>% 
  drop_na() %>% 
  summarise(median = median(flipper_length_mm), 
            sd = sd(flipper_length_mm)) %>% 
  mutate(across(where(is.numeric), ~ round(., 1)))

Wide vs long format

Load data

gap_wide <- read_csv("https://bit.ly/gapminder-rsu")

## Rows: 142 Columns: 38

## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr  (2): continent, country
## dbl (36): gdpPercap_1952, gdpPercap_1957, gdpPercap_1962, gdpPercap_1967, gd...

## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

Explore the data

How many variables?

Filter China

How to plot ?

Wide vs long

Wide data is for humans
Long data is for computers

Wide to long

gdp

gap_wide %>% 
  select(continent:gdpPercap_2007) %>% 
  pivot_longer(cols = gdpPercap_1952:gdpPercap_2007, 
               names_to = "gdp_year", 
               values_to = "gdp_value") %>% 
  filter(country == "China") %>% 
  ggplot(aes(x = gdp_year, 
             y = gdp_value, 
             group = country)) + 
  geom_line()

try to plot

Filter only China

Conects the points

Filter Europe and conects the points

Split-apply-combine

Now, we will use the long gapminder dataset

Question: how has GDP per continent evolved per year?

gapminder %>% 
  filter(year == 2007) %>% 
  group_by(continent) %>% 
  summarise(mean_gdp = mean(gdpPercap), 
            mean_pop = mean(pop), 
            mean_lifeexp = mean(lifeExp), 
            n = n())

gapminder %>% 
  group_by(continent, year) %>% 
  summarise(gdp_mean = mean(gdpPercap)) %>% 
  ggplot(aes(x = year, 
             y = gdp_mean, 
             color = continent)) + 
  geom_line() + 
  scale_y_log10()

## `summarise()` has grouped output by 'continent'. You can override using the `.groups` argument.

Question: how has POPULATION per continent evolved per year?

Hint: Try log10

Example: Vaccinations Latvia

Homework: See <https://github.com/owid/covid-19-data/tree/master/public/data/vaccinations>

covid_vac <- read_csv("https://github.com/owid/covid-19-data/raw/master/public/data/vaccinations/vaccinations.csv")

## Rows: 63861 Columns: 16

## ── Column specification ────────────────────────────────────────────────────────
## Delimiter: ","
## chr   (2): location, iso_code
## dbl  (13): total_vaccinations, people_vaccinated, people_fully_vaccinated, t...
## date  (1): date

## 
## ℹ Use `spec()` to retrieve the full column specification for this data.
## ℹ Specify the column types or set `show_col_types = FALSE` to quiet this message.

covid_vac %>% 
  filter(location == "Latvia") %>% 
  ggplot(aes(x = date, 
             y = total_vaccinations)) + 
  geom_point()

## Warning: Removed 10 rows containing missing values (geom_point).

2021 4th