Some useful R commands.

Useful_R_Commands.md

##1 Data manipulation

1. If the data contains NA values, it regards it as factor, not numeric.

    DATA$COLUMN <- as.numeric(as.character(DATA$COLUMN))
   

2. Rename the column:

    names(DATA)[2] <- "NEW_NAME"
   

3. Melt the data for ggplot2 plotting

    df <- melt(DATA, id.vars="COLUMN", na.rm=T)
   

4. Remove NA in melt:

    na.rm = T
   

5. Remove NA from a dataframe

    df <- na.omit(df)
   

6. Change the digits of data

    format(DATA$COLUMN, digits=4, scientific=F)
   

7. Change the dimnames of a matrix

    dimnames = list(c("O", "C", "L", "H"), c("O", "C", "L", "H"))
   

8. Extract the specific area of a data frame

    DATA2 <- DATA[1:(nrow(DATA)-1), 3:5] 
   
    #extract the value from 1st row to (nrow-1), 3rd-5th columns
   

9. Add the ratio of two specific colomn to a data frame

    ratio <- transform(DATA, ratio = paste(round(COLUMN1 / COLUMN2*100, digits=2), "%")
   

10. Remove rows from a dataframe according to specific conditions

    a <- a[which(a$COL1!= "2012-03-24" & a$COL1!= "2012-03-25"] #diffrent rows
    
    or
    
    a <- a[!(a$COL2=="A" & a$COL3=="B"),] #one row with different conditions
    

##2 ggplot2

1. Change the labels of xlab

    scale_x_date(labels = date_format("%m-%Y")) #need timeDate package
   

2. Omit the title of xlab

    xlab("")
   

3. Change the color to grey scale

    scale_color_grey()
   

4. Change the position of legend

    theme(legend.position="bottom")
   

5. Annotate the plot

    myvalue <- format(DATA$COL, digits=4, scientific=F) #Set the tag
   
    geom_text(aes(Var1, Var2, label=myvalue), color="snow")
   

6. Change the font size of axes

    theme(axis.text.x=element_text(size=15), axis.text.y=element_text(size=15))
   

7. Remove the legend

    theme(legend.position="none")
   

8. Change the legend label and change the color to gray scale

    scale_fill_grey(labels=c("A", "B")
   

9. Print a high-res png file

    ggsave(plot=last_plot(), file="FILENAME.png", dpi=300)
   

10. Set the limit of axix

    scale_y_continous(limits=(0, 200))
    
    xlim(0,30)
    

##3 ACF and CCF test

1. Manipulate the data to a matrix

    m.DATA <- as.matrix(DATA[2:5])
   

2. Apply the ACF test

    acf.DATA <- acf(DATA, na.action=na.contiguous, plot = F) 
    
    #find the longest contiguous stretch of non-NAs
    #turn off the plot
   

3. Extract the ACF matrix from the ACF result

    m.acf.DATA <- matrix(acf.DATA$acf, ncol = 4, byrow = T)
   
    #make a matrix according to the result
   

4. Change the label of the axis in heatmap of ACF result

    DATA$COLUMN[DATA$COLUMN == 1] <- "Day0"
   

Machine learning

rpart(y~., data)
rpart(y~.-V1, data)

library(e1071) 
svm(x=xMatrix, y=yVector, ...) 


xnam <- paste("x", c(1,3,9:20,50), sep="") 
fmla <- as.formula(paste("y ~ ", paste(xnam, collapse= "+")))