d <- read.csv(file = 'https://stats.dip.jp/01_ds/data/data_cleansing2_dirty_data_exercise.csv')

library(DT)
datatable(d)

m <- as.matrix(d) 
boxplot(m)

any(m < 0)   # 負のデータが存在

## [1] TRUE

any(abs(m) > 3*sd(m)) # 3シグマ（標準偏差の３倍）を超えるデータが存在

## [1] TRUE

jj <- 25:85
boxplot(m[, jj])

j <- 27
y <- m[, j]

barplot(height    = y,         
        names.arg = 1:nrow(m))

ii <- 500:550
y <- m[ii, j]

barplot(height    = y,                        
        cex.names = 0.4,                      
        names.arg = paste(ii, '\n(', y, ')'))

ii <- 527:535
y <- m[ii, j]

barplot(height    = y,                        
        cex.names = 0.4,                      
        names.arg = paste(ii, '\n(', y, ')'))

m[531, j]

##   V27 
## -0.92

m[m < 0]

## [1] -42.00  -0.92  -0.55  -0.03

sigma <- sd(m)
m[abs(m) > 4*sigma]

## [1] -42  22 111

m.a <- NULL 

for (j in 1:ncol(m))
{
  for (i in 1:nrow(m))
  { 
    if ( m[i, j] < 0 | m[i, j] > 4 * sigma )
    {
      cat('Row:', i, ', Col:', j, ', Value: ', m[i, j], fill = T) 
      m.a <- rbind(m.a, t(c(i, j, m[i, j]))) 
    }
  }
}

## Row: 2 , Col: 1 , Value:  -42
## Row: 42 , Col: 1 , Value:  22
## Row: 99 , Col: 1 , Value:  111
## Row: 531 , Col: 27 , Value:  -0.92
## Row: 675 , Col: 83 , Value:  -0.55
## Row: 124 , Col: 85 , Value:  -0.03

colnames(m.a) <- c('Row', 'Col', 'Value')
m.a

##      Row Col  Value
## [1,]   2   1 -42.00
## [2,]  42   1  22.00
## [3,]  99   1 111.00
## [4,] 531  27  -0.92
## [5,] 675  83  -0.55
## [6,] 124  85  -0.03