Example Data

During the courses, example data files are used to illustrate different concepts. These files are available here for download or for direct usage.

Loading the data files

Each data file can be separately loaded from the URLs reported down below. Nonetheless, you can also exploit the ability of R read.* functions to open a file directly from its URL. To do so, it can be useful to use the adas.utils::examples_url() function that builds the URL based on the file name:

examples_url("timeseries.csv")

[1] "https://paolobosetti.quarto.pub/data/timeseries.csv"

Then, just pipe it with the proper file-reading function, e.g.:

examples_url("timeseries.csv") %>% read_csv()

or

examples_url("3dprint.dat") %>% read_table(comment="#")

List of CSV files

• duplicate.csv
• kfold.csv
• prod_defects.csv
• regression.csv
• surface_scan.csv
• taratura.csv
• temperature-anomaly.csv
• timeseries.csv
• train.csv

List of DAT files

• 3dprint.dat
• anova.dat
• battery.dat
• bearing.dat
• besterfield_1.dat
• besterfield_2.dat
• besterfield_3.dat
• besterfield_4.dat
• besterfield_5.dat
• cotton.dat
• cutting.dat
• diet.dat
• drill.dat
• hardness.dat
• hardness2.dat
• mortar.dat
• p_data.dat
• rf_generator.dat
• soap_bottles.dat
• twosample.dat

Example

Let’s look at the train.csv file from the Titanic dataset. We load the dataset and perform some basic data manipulation to only keep the relevant columns. We also classify the passengers into age classes.

titanic <- examples_url("train.csv") %>% 
  read_csv(show_col_types = FALSE) %>% 
  mutate(
    Survived = as.logical(Survived),
    Pclass = as.factor(Pclass),
    AgeClass = cut(Age, breaks=c(0, 10, 18, 40, 60, Inf), labels=c("child", "young", "adult", "senior", "elderly")),
    Age = as.factor(Age)
  ) %>% 
  select(
    PassengerId,
    Survived,
    Pclass,
    Sex,
    Age,
    AgeClass,
    Fare
  )
titanic %>% head() %>% knitr::kable()

PassengerId	Survived	Pclass	Sex	Age	AgeClass	Fare
1	FALSE	3	male	22	adult	7.2500
2	TRUE	1	female	38	adult	71.2833
3	TRUE	3	female	26	adult	7.9250
4	TRUE	1	female	35	adult	53.1000
5	FALSE	3	male	35	adult	8.0500
6	FALSE	3	male	NA	NA	8.4583

Let’s plot the survival rate by age and sex class, as a histogram.

titanic %>% 
  ggplot(aes(x=AgeClass, fill=Survived)) +
  geom_bar() +
  facet_wrap(~Sex) +
  labs(title="Survival by Age Class", x="Age Class", y="Count", fill="Survived")

In a more refined way, we can build a generalized linear model of binomial type🇮🇹 to predict the survival rate based on the fare, age class, and sex:

model <- glm(Survived ~ Fare + AgeClass + Sex, data=titanic, family=binomial)

pred <- expand.grid(
  Fare = seq(0, 500, 5),
  AgeClass = levels(titanic$AgeClass),
  Sex = c("female", "male")
) %>% 
  add_predictions(model, var="pred.glm", type="response")

titanic %>%
  add_predictions(model, var="pred.glm", type="response") %>% 
  ggplot(aes(x=Fare, y=pred.glm, color=AgeClass)) +
  geom_line(data=pred) +
  geom_point(aes(shape=Survived), size=2) +
  coord_cartesian(ylim=c(0, 1)) +
  facet_wrap(~Sex) +
  scale_shape_manual(values=c(3, 1)) +
  labs(y="Predicted Survival", color="Age Class")

A word to the wise

According to our analysis, if you embark the Titanic, you better be a young and rich female 😉.