crime_data_reformatter.r

#===============================================================================
#                                                                       TOY DATA
#===============================================================================
# Setting up a toy dataframe
df = data.frame(total_crime_2015_2016=674, 
                total_crime_2014_2015=323, 
                total_crime_2013_2014=212, 
                car_theft_2015_2016=34,
                car_theft_2014_2015=45,  
                car_theft_2013_2014=74
            )
row.names(df) = "count"


#===============================================================================
#                                                                       GET_YEAR
#===============================================================================
#' get_year
#' 
#' Takes a string `s` that takes the form such as: 
#' 
#'      "total_crime_2013_2014"
#'      
#' And returns just the year range component of that string, eg:
#' 
#'      "2013_2014"
#'      
#' @param s (string) the string to process
get_year <- function(s){
    #  .*                   = any number of non-newline characters
    #  ()                   = Anything enclosed in this, will specify a pattern 
    #                         that we can refer to later on. 
    #  [0-9]{4}_[0-9]{4}    = exactly 4 digits between 0 and 9, followed by an 
    #                         underscore, followed by exactly 4 more digits. 
    #  $                    = specifies that the last pattern should be located 
    #                         at the very end of the string. 
    #
    # "\\1"                 = gsub  looks for the above pattern we specified 
    #                         within the string s, and replaces it with the 
    #                         first pattern that we enclosed within round 
    #                         brackets
    return(gsub(".*([0-9]{4}_[0-9]{4})$", "\\1", s))
}




#===============================================================================
#                                                    Extract Columns of Interest
#===============================================================================
# extract values from columns that satisfy some regular expression pattern. 
# eg: 
#    ^total_crime_[0-9]{4} tells us that we are : 
#      - looking for strings that START with "total_crime_"
#      - that it should then be followed by 4 numbers
# This should be enough to uniquely pick out the relevant fields without overlap
 
total_crime = df[, grep("^total_crime_[0-9]{4}", names(df), ignore.case=FALSE)]
row.names(total_crime) = "total_crime"

car_theft   = df[, grep("^car_theft_[0-9]{4}", names(df), ignore.case=FALSE)]
row.names(car_theft) = "car_theft"

#===============================================================================
#                                                        Rename the Column Names
#===============================================================================
# Rename the columns to only contain the years
names(total_crime) = sapply(names(total_crime), get_year)
names(car_theft) = sapply(names(car_theft), get_year)

#===============================================================================
#                                                       Create the new DataFrame
#===============================================================================
# Create a new dataframe with the extracted data
new_df = rbind(total_crime, car_theft)


# Creates a dataframe that looks like this
#             2015_2016 2014_2015 2013_2014
# total_crime       674       323       212
# car_theft          34        45        74