Stats descr.

Statistique descriptive : résumer l’information avec caractéristiques essentielles.

Comment ? Avec des indicateurs et des représentations graphiques.

Stats descr.

bdd <- read.table("./DATA/bdd.csv", header = TRUE, sep = ",", dec = ".")
head(bdd)

##   year month day hour minute second temperature      gas humidity pressure
## 1 2018    10  22   17     58      1    10.75793 11811196 80.34636 1040.177
## 2 2018    10  22   17     59     26    10.76887 11682953 79.98910 1040.153
## 3 2018    10  22   18      0     29    10.80402 11569892 79.74389 1040.090
## 4 2018    10  22   18      1     31    10.78391 11582347 79.84094 1040.164
## 5 2018    10  22   18      2     33    10.73586 11582347 80.57278 1040.286
## 6 2018    10  22   18      3     36    10.66633 11582347 81.02920 1040.477
##   lightVisible lightIR lightUV
## 1          264     318       4
## 2          266     313       5
## 3          263     306       3
## 4          267     307       4
## 5          266     308       5
## 6          266     307       5

Stats descr.

nrow(bdd)

## [1] 465540

ncol(bdd)

## [1] 13

Stats descr.

# moyenne
mean(bdd$temperature)

## [1] 11.12679

# médiane
median(bdd$temperature)

## [1] 9.93

# minimum
min(bdd$temperature)

## [1] -5.986406

# maximum
max(bdd$temperature)

## [1] 49.01672

Stats descr.

# sqrt(variance) = écart-type
n <- length(bdd$temperature)
xi <- bdd$temperature
sn <- 1/(n-1) * sum(xi^2 - mean(xi)^2)
print(sn)

## [1] 73.94156

var(xi) # avec R

## [1] 73.94156

sqrt(sn)

## [1] 8.598928

sd(xi) # avec R

## [1] 8.598928

Stats descr.

# quantiles
quantile(xi)

##        0%       25%       50%       75%      100% 
## -5.986406  4.618672  9.930000 16.842305 49.016719

quantile(xi, probs = 0.5)

##  50% 
## 9.93

Stats descr.

summary(bdd$temperature)

##    Min. 1st Qu.  Median    Mean 3rd Qu.    Max. 
##  -5.986   4.619   9.930  11.127  16.842  49.017

Stats descr.

summary(bdd[c("temperature", "humidity", "pressure")])

##   temperature        humidity         pressure     
##  Min.   :-5.986   Min.   : 15.27   Min.   : 980.6  
##  1st Qu.: 4.619   1st Qu.: 54.70   1st Qu.:1011.2  
##  Median : 9.930   Median : 74.09   Median :1025.5  
##  Mean   :11.127   Mean   : 69.41   Mean   :1027.4  
##  3rd Qu.:16.842   3rd Qu.: 85.23   3rd Qu.:1041.5  
##  Max.   :49.017   Max.   :100.00   Max.   :1085.6

Stats descr.

summary(bdd[c(7, 9, 10)])

##   temperature        humidity         pressure     
##  Min.   :-5.986   Min.   : 15.27   Min.   : 980.6  
##  1st Qu.: 4.619   1st Qu.: 54.70   1st Qu.:1011.2  
##  Median : 9.930   Median : 74.09   Median :1025.5  
##  Mean   :11.127   Mean   : 69.41   Mean   :1027.4  
##  3rd Qu.:16.842   3rd Qu.: 85.23   3rd Qu.:1041.5  
##  Max.   :49.017   Max.   :100.00   Max.   :1085.6

Stats descr.

moyennes par mois ?

bdd$date <- as.POSIXct(
  paste(
    bdd$year, bdd$month, bdd$day, bdd$hour, bdd$minute, bdd$second),
  format = "%Y %m %d %H %M %S"
)
head(bdd$date)

## [1] "2018-10-22 17:58:01 CEST" "2018-10-22 17:59:26 CEST"
## [3] "2018-10-22 18:00:29 CEST" "2018-10-22 18:01:31 CEST"
## [5] "2018-10-22 18:02:33 CEST" "2018-10-22 18:03:36 CEST"

Stats descr.

dateDay <- format(bdd$date, "%Y/%m/%d")
tempDay <- tapply(
  X = bdd$temperature, 
  INDEX = dateDay,
  FUN = mean)

Stats descr.

plot(
  x = as.Date(names(tempDay), format = "%Y/%m/%d"),
  y = tempDay, xlab = "Time",ylab = "Temperature (°C)",
  type = "l", lwd = 2)

Stats descr.

temp3q <- tapply(
  X = bdd$temperature, 
  INDEX = dateDay,
  FUN = quantile, probs = 0.75)
temp1q <- tapply(
  X = bdd$temperature, 
  INDEX = dateDay,
  FUN = quantile, probs = 0.25)

Stats descr.

plot(
  x = as.Date(names(tempDay), format = "%Y/%m/%d"),
  y = tempDay, xlab = "Time",ylab = "Temperature (°C)",
  type = "l", lwd = 2)
points(
  x = as.Date(names(temp3q), format = "%Y/%m/%d"),
  y = temp3q, xlab = "Time",ylab = "Temperature (°C)",
  type = "l", lwd = 2, col = "red")
points(
  x = as.Date(names(temp1q), format = "%Y/%m/%d"),
  y = temp1q, xlab = "Time",ylab = "Temperature (°C)",
  type = "l", lwd = 2, col = "blue")

Stats descr.

Stats descr.

boxplot(bdd$temperature ~ format(bdd$date, "%Y/%m/%d"))

Stats descr.

par(mar = c(6, 2, 1, 1))
boxplot(bdd$temperature ~ format(bdd$date, "%Y/%U"), 
  las = 3, xlab = "", ylab = "")

Stats descr.

par(mar = c(6, 2, 1, 1))
boxplot(bdd$temperature ~ format(bdd$date, "%Y/%m"), 
  las = 3, xlab = "", ylab = "")

Stats descr.

tempMonthMean <- tapply(bdd$temp, 
  INDEX = format(bdd$date, format = "%Y-%m"), FUN = mean)
myCol <- colorRampPalette(c("blue", "red"))(101)
tempMeanDayPos <- round(
  (tempMonthMean - min(tempMonthMean)) / 
    (max(tempMonthMean) - min(tempMonthMean))*100) + 1
par(mar = c(6, 2, 1, 1))
boxplot(bdd$temp ~ format(bdd$date, format = "%Y-%m"), las = 3, 
  col = myCol[tempMeanDayPos],
  xlab = "", ylab = "")

Stats descr.

Stats descr.

# https://fr.climate-data.org
tempRefMoy <- c(11.4, 6.8, 3.9, 3, 3.9, 7.6, 10.6, 14, 17.1, 
  19, 18.7, 16.2)
tempRefMin <- c(7.1, 3.7, 1.3, 0.1, 0.5, 2.8, 5.3, 8.6, 11.6, 
  13.5, 13.3, 11.1)
tempRefMax <- c(15.7, 9.9, 6.5, 5.9, 7.3, 12.4, 15.9, 19.5, 
  22.7, 24.6, 24.2, 21.3)

Stats descr.

boxplot(bdd$temp ~ format(bdd$date, format = "%Y-%m"), las = 3, 
  col = myCol[tempMeanDayPos],
  xlab = "", ylab = "", outline = FALSE)
points(tempRefMoy, type = 'l', lwd = 2)
points(tempRefMin, type = 'l', lwd = 2, col = 4)
points(tempRefMax, type = 'l', lwd = 2, col = 2)

Stats descr.

Stats descr.

Nombre de jours avec une temperature moyenne supérieure à 30°C ? Nombre de jours avec une temperature moyenne inférieure à 5°C ?

Stats descr.

length(tempDay[tempDay >= 30])

## [1] 2

length(tempDay[tempDay <= 5])

## [1] 74

Stats descr.

La température est une variable continue :

head(bdd$temperature)

## [1] 10.75793 10.76887 10.80402 10.78391 10.73586 10.66633

Stats descr.

myCut <- seq(from = -10, to = 60, by = 2)
print(myCut)

##  [1] -10  -8  -6  -4  -2   0   2   4   6   8  10  12  14  16  18  20  22
## [18]  24  26  28  30  32  34  36  38  40  42  44  46  48  50  52  54  56
## [35]  58  60

bddCut <- cut(bdd$temperature, breaks = myCut)
head(bddCut)

## [1] (10,12] (10,12] (10,12] (10,12] (10,12] (10,12]
## 35 Levels: (-10,-8] (-8,-6] (-6,-4] (-4,-2] (-2,0] (0,2] (2,4] ... (58,60]

Stats descr.

tCut <- table(bddCut)
print(tCut)

## bddCut
## (-10,-8]  (-8,-6]  (-6,-4]  (-4,-2]   (-2,0]    (0,2]    (2,4]    (4,6] 
##        0        0     3366    12700    21865    28247    36015    45741 
##    (6,8]   (8,10]  (10,12]  (12,14]  (14,16]  (16,18]  (18,20]  (20,22] 
##    41689    44379    36419    34764    30073    30975    23612    19355 
##  (22,24]  (24,26]  (26,28]  (28,30]  (30,32]  (32,34]  (34,36]  (36,38] 
##    17278    12372     9352     6044     4822     2256     1618     1292 
##  (38,40]  (40,42]  (42,44]  (44,46]  (46,48]  (48,50]  (50,52]  (52,54] 
##      699      322      110       73       71       31        0        0 
##  (54,56]  (56,58]  (58,60] 
##        0        0        0

Stats descr.

barplot(tCut, las = 3)

Stats descr.

barplot(
  table(
    cut(
      bdd$temperature, 
      breaks = seq(from = -10, to = 60, by = 2)
    )
  ), las = 3
)

Stats descr.

hist(bdd$temperature, breaks = myCut)

Stats descr.

hist(bdd$temperature, 
  breaks = c(-10, 0, 5, 10, 15, 20, 30, 100))

Stats descr.

Fn(x) : % obs < x

plot(ecdf(bdd$temperature))

Inkscape

https://inkscape.org

Inkscape

barplot(sample(10:15, 8, replace = TRUE), 
  col = 1:8, names.arg = 1:8)

Inkscape

pdf("barplot.pdf")
  barplot(sample(10:15, 8, replace = TRUE), 
    col = 1:8, names.arg = 1:8)
dev.off()

## png 
##   2

Inkscape

LINK VIDEO

Inkscape

pdf("multplot.pdf")
  boxplot(rnorm(100), rnorm(100))
  plot(rnorm(100))
  hist(rnorm(100))
dev.off()

## png 
##   2

Inkscape

The Gimp

https://www.gimp.org

The Gimp

File / Export as… / tif / LZW compression