Analyzing data by group

$$ When you$$re analyzing a data set, often you want to

calculate the metrics broken down by each category of

another variable.

$$ In the previous homework, you wrote a function

$($generateNumericSummary$)$ to summarize a numeric

variable by another categorical variable.

$$ By generalizing$/$extending the previous code, we will

design an data object to analyze a variable by another

categorical variable.

Class dataByGroup

$$ dataByGroup $($abstract class$)$

$$ Contains two pandas$$ series: dat and group

$$ numericDataByGroup

$$ Sub$-$class of dataByGroup

$$ dat should be numeric.

$$ categoricalDataByGroup

$$ Sub$-$class of dataByGroup

$$ dat should be categorical.

dataByGroupdataByGroup

numericDataByGroupnumericDataByGroup categoricalDataByGroupcategoricalDataByGroup

Class dataByGroup

$$ Write a class definition for object dataByGroup with the

following specification:

$$ Object attributes

$$ dat: $($pandas$)$ Series

$$ group: categorical $($pandas$)$ Series of the same length as dat

$\_$isBinary: Boolean $($True$/$False$),$ indicating whether dat is binary.

$$ methods

$\_\_$init$\_\_($self$,$ dat, group$)$: takes dat and group as input data and initialize dat and

group of object dataByGroup.

$\_\_$str$\_\_($self$)$: print dat and group $($combined together$)$ as a $($pandas$)$ DataFrame

$$ isBinary$($self$)$: returns the value of $\_$isBinary. $($Accessor method$)$

$$ getNumMissings$($self$)$: returns the number of missing values in dat. $($Accessor

method$)$

Class dataByGroup

$$ Class header

import pandas as pd

import numpy as np

class dataByGroup$($object$)$:

def $\_\_$init$\_\_($self$,$ dat, group$)$:

#write your code

# $\_\_$init$\_\_$ should initialize three object attributes

# dat, group, $\_$isBinary

def $\_\_$str$\_\_($self$)$:

#write your code

def isBinary$($self$)$:

#write your code

def getNumMissings$($self$)$:

#write your code

Sub$-$Class numericDataByGroup

$$ Write a class definition for object numericDataByGroup

with the following specification:

$$ Object attributes

$$ Inherits all attributes of its super$-$class $-$ dataByGroup

$$ methods

$$ Inherits all methods of its super$-$class $-$ dataByGroup

$$ getMeans$($self$)$: returns means of dat across the different levels of

group.

$$ Return data type: pandas Series

$$ getSTD$($self$)$: returns standard deviations of dat across the different

levels of group.

$$ Return data type: pandas Series

Sub$-$Class numericDataByGroup

$$ Class header

class numericDataByGroup$($dataByGroup$)$:

def $\_\_$init$\_\_($self$,$dat, group$)$:

#write your code

def getMeans$($self$)$:

#write your code

def getSTD$($self$)$:

#write your code

Sub$-$Class categoricalDataByGroup

$$ Write a class definition for object

categoricalDataByGroup with the following

specification:

$$ Object attributes

$$ Inherits all attributes of its super$-$class $-$ dataByGroup

$$ Methods

$$ Inherits all methods of its super$-$class $-$ dataByGroup

$$ getTallies$($self$)$: returns tabulated counts $($tallies$)$ by dat and group

$$ Return data type: pandas Series

Sub$-$Class categoricalDataByGroup

$$ Class header

class categoricalDataByGroup$($dataByGroup$)$:

def $\_\_$init$\_\_($self$,$dat, group$)$:

#write your code

def getTallies$($self$)$:

#write your code

Output

$$ Test case $1$: categorical data

def main$()$:

titanic $=$ pd$.$read$\_$csv$("$titanic$3.$csv$",$header$=0)$

survivedByPclass $=$categoricalDataByGroup$($titanic$[\text{'}$survived$\text{'}],$titanic$[\text{'}$pclass$\text{'}])$

print$("$Data and Group: $")$

print$($survivedByPclass$)$ ## $\_\_$str$\_\_$ is invoked

print$("$Is the data binary? : $"+$str$($survivedByPclass$.$isBinary$()))$

print$("$The number of missing values : $"+$str$($survivedByPclass$.$getNumMissings$()))$

print$("$Tallies: $")$

print$($survivedByPclass$.$getTallies$())$

Test case $2$: numerical data

def main$()$:

ageBySurvived $=$ numericDataByGroup$($titanic$[\text{'}$age$\text{'}],$titanic$[\text{'}$survived$\text{'}])$

print$("$Data and Group: $")$

print$($ageBySurvived$)$ ## $\_\_$str$\_\_$ is invoked

print$("$Is the data binary? : $"+$str$($ageBySurvived$.$isBinary$()))$

print$("$The number of missing values : $"+$str$($ageBySurvived$.$getNumMissings$()))$

print$("$Means: $")$

print$($ageBySurvived$.$getMeans$())$

print$("$Standard Deviations: $")$

print$($ageBySurvived$.$getSTD$())$

Here is my code:

import numpy as np

import pandas as pd

class dataByGroup$($object$)$:

def $\_\_$init$\_\_($self$,$ dat, group$)$:

self.dat $=$ dat

self.group $=$ group

def $\_\_$str$\_\_($self$)$:

df$\_$dat $=$ pd$.$DataFrame$(\{$self$.$dat.name: self.dat$\})$

df$\_$group $=$ pd$.$DataFrame$(\{$self$.$group.name: self.group$\})$

return df$\_$merged.to$\_$string$()$

def isBinary$($self$)$:

lst $=$ np$.$array$($self$.$dat$)$

lst$\_$check $=$ np$.$logical$\_$or$($lst $==1,$ lst $==0)$

lst$\_$check$\_$nan $=$ np$.$isnan$($lst$)$

return $($lst$\_$final$\_$check $==$ True$).$all$()$

def getNumMissings$($self$)$:

isnull $=0$

check$\_$null$\_$df $=$ pd$.$isnull$($self$.$dat$)$

for i in range$($len$($self$.$dat$))$:

if check$\_$null$\_$df$[$i$]==$ True:

isnull $+=1$

return isnull

class numericDataByGroup$($dataByGroup$)$:

def $\_\_$init$\_\_($self$,$ dat, group$)$:

dataByGroup.$\_\_$init$\_\_($self$,$ dat, group$)$

def getMeans$($self$)$:

return self.dat.groupby$($self$.$group$).$mean$()$

def getSTD$($self$)$:

return self.dat.groupby$($self$.$group$).$std$()$