Slides, script, données

01_Intro/EDA.R

@@ -1,2 +1,107 @@
 library(tidyverse)
+library(gapminder)
+library(lubridate)
 
+# Preliminaire ----
+
+# pipe : %>%
+# Permet de chaîner les instructions
+# x %>% f         <=> f(x)
+# x %>% f(.)      <=> f(x)
+# x %>% f(y = 10) <=> f(x, y = 10)
+
+# EDA ----
+
+## Diamonds ----
+
+data(diamonds)
+
+diamonds
+
+diamonds %>%
+  str
+
+diamonds %>%
+  summary
+
+diamonds %>%
+  keep(is.numeric) %>%
+  map(sd)
+
+diamonds %>%
+  ggplot() +
+  aes(x = carat) +
+  geom_histogram(binwidth = .05)
+
+diamonds %>%
+  ggplot() +
+  aes(x = cut) +
+  geom_bar()
+
+# …
+
+plot_desc <- function(x, var, ...)
+{
+  x %>%
+    ggplot() +
+    aes_string(x = var) -> g
+
+    if (x[[var]] %>% is.numeric)
+      g + geom_histogram(...)
+    else
+      g + geom_bar(...)
+}
+
+diamonds %>%
+  plot_desc("carat", binwidth = .01)
+
+diamonds %>%
+  keep(is.numeric) %>%
+  names %>%
+  map(~plot_desc(diamonds, ., binwidth = .5)) %>%
+  set_names(diamonds %>% keep(is.numeric) %>% names) -> plots
+
+diamonds %>%
+  keep(is.factor) %>%
+  names %>%
+  map(~plot_desc(diamonds, .)) %>%
+  set_names(diamonds %>% keep(is.factor) %>% names) -> plots
+
+diamonds %>%
+  sample_n(1000) %>%
+  plot
+
+diamonds %>%
+  ggplot() +
+  aes(x = carat, fill = clarity) +
+  geom_histogram(position = "fill")
+
+diamonds %>%
+  ggplot() +
+  aes(x = carat, fill = cut) +
+  geom_histogram(position = "fill")
+
+diamonds %>%
+  ggplot() +
+  aes(x = carat, fill = color) +
+  geom_histogram(position = "fill")
+
+diamonds %>%
+  ggplot() +
+  aes(x = x, y = y) +
+  geom_point()
+
+diamonds %>%
+  ggplot() +
+  aes(x = carat, y = price, color = color) +
+  geom_point()
+
+## Exercice gapminder ----
+
+data(gapminder)
+
+# Feature engineering ----
+
+read_csv2("commits.csv")
+
+# Commits par jour ? En fonction de l'heure ? Par jour et par heure ? boulot/perso ? en semaine/week-end ?