import pandas as pd

import numpy as np

#$\%$matplotlib inline

#import pandas$\_$profiling

import seaborn as sns

from sklearn.metrics import roc$\_$auc$\_$score

amazonData$=$pd$.$read$\_$csv$("$Amazon$\_$Unlocked$\_$Mobile.csv$")$

amazonData.head$(5)$

amazonData.describe$()$

amazonData.info$()$

#plot to see how rating is distributed

sns$.$countplot$($amazonData$[\text{'}$Rating$\text{'}])$

#Data cleaning

amazonData.drop$\_$duplicates$($keep$=$False,inplace$=$True$)$

amazonData$=$amazonData.dropna$()$

rowcount$=$amazonData.shape$[0]$

rowcount

#relation between price and rating

#sns$.$regplot$($x$=$'Price',y$=$'Rating',data$=$amazonData$)$

amazonData.isnull$().$sum$()$

amazonData$["$Reviews$"].$head$(2)$

amazonData$["$Reviews$"][0]$

sentiData$=$amazonData$[["$Reviews$",$"Rating"$]]$

sentiData$=$sentiData$.$iloc$[$:$20000]$

sentiData$.$head$(5)$

sentiData$["$Reviews$"][1]$

#distinct products

amazonData$["$Product Name"$].$nunique$()$

#distinct brand

amazonData$["$Brand Name"$].$nunique$()$

#Top $10$ Brand Names

pivot$1=$ pd$.$pivot$\_$table$($amazonData$,$

values $=[\text{'}$Rating$\text{'},$ 'Review Votes'$],$

index $=[\text{'}$Brand Name'$],$

columns$=[],$

aggfunc$=[$np$.$sum, np$.$mean, np$.$count$\_$nonzero$],$

margins$=$True, fill$\_$value$=0).$sort$\_$values$($by$=(\text{'}$count$\_$nonzero', 'Rating'$),$ ascending$=$False$).$fillna$(\text{'}\text{'})$

topmost$\_$prods $=$ pivot$1.$reindex$().$head$($n$=10)$

topmost$\_$prods $=$ topmost$\_$prods.reset$\_$index$()$

topmost$\_$prods

###### Top $10$ products in Samsung

df$\_$samsung $=$ amazonData.loc$[$amazonData$[\text{'}$Brand Name'$].$isin$([\text{'}$Samsung$\text{'}])]$

pivot $=$ pd$.$pivot$\_$table$($df$\_$samsung,

values $=[\text{'}$Rating$\text{'},$ 'Review Votes'$],$

index $=[\text{'}$Product Name'$],$

columns$=[],$

aggfunc$=[$np$.$sum, np$.$mean, np$.$count$\_$nonzero$],$

margins$=$True, fill$\_$value$=0).$sort$\_$values$($by$=(\text{'}$count$\_$nonzero', 'Rating'$),$ ascending$=$False$).$fillna$(\text{'}\text{'})$

topmost$\_$prods$\_$samsung $=$ pivot.reindex$().$head$($n$=10)$

topmost$\_$prods$\_$samsung $=$ topmost$\_$prods$\_$samsung.reset$\_$index$()$

topmost$\_$prods$\_$samsung

###### Top $10$ products in apple

df$\_$apple $=$ amazonData.loc$[$amazonData$[\text{'}$Brand Name'$].$isin$([\text{'}$Apple$\text{'}])]$

pivot $=$ pd$.$pivot$\_$table$($df$\_$apple,

values $=[\text{'}$Rating$\text{'},$ 'Review Votes'$],$

index $=[\text{'}$Product Name'$],$

columns$=[],$

aggfunc$=[$np$.$sum, np$.$mean, np$.$count$\_$nonzero$],$

margins$=$True, fill$\_$value$=0).$sort$\_$values$($by$=(\text{'}$count$\_$nonzero', 'Rating'$),$ ascending$=$False$).$fillna$(\text{'}\text{'})$

topmost$\_$prods$\_$apple $=$ pivot.reindex$().$head$($n$=10)$

topmost$\_$prods$\_$apple $=$ topmost$\_$prods$\_$apple.reset$\_$index$()$

topmost$\_$prods$\_$apple

#installing NLTK package