Whatsapp Tools

install

pip install -e git+https://github.com/gauravmeena0708/whatsapp-tools#egg=whatsapp-tools
pip install emoji wordcloud

Example Use

from whatsapptools import GroupAnalyzer
analyzer = GroupAnalyzer("data/WhatsApp_Chat3.txt")
df = analyzer.parse_chat_data()
df.head()

df1 = analyzer.df_basic_cleanup(df)
df1.head()

from nltk.corpus import stopwords

custom_stopwords = {"Bhai", "<Media omitted>", "Media omitted", "Media","omitted", "bro", 'would', 'ye', 'ke', 'ko', 'doge', 'aap', 'tum', 'hai'
                    }
stop_words1 = set(stopwords.words('english'))
stop_words2 = set(stopwords.words('english')).union(custom_stopwords)
analyzer.generate_wordcloud(df["message"].str.cat(sep=" "), stop_words1)
analyzer.generate_wordcloud(df["message"].str.cat(sep=" "), stop_words2)

analyzer.create_plotly_fig(df,'name','message','message')
analyzer.create_plotly_fig(df,'ym','message','ym',True)
#create_plotly_fig(df,'dt','message','dt',True)
analyzer.create_plotly_fig(df,'day','message','day',False)
analyzer.create_plotly_fig(df,'yd','message','yd',True)
analyzer.create_plotly_fig(df,'name','emojicount','emojicount',True,False)
analyzer.create_plotly_fig(df,'name','urlcount','urlcount',True,False)
analyzer.create_plotly_fig(df1,'name','emojicount','emojicount')
analyzer.create_plotly_fig(df1,'name','urlcount','urlcount')
analyzer.create_plotly_fig(df1,'name','yturlcount','yturlcount')
analyzer.create_plotly_fig(df1,'name','mediacount','mediacount',count=False)
analyzer.create_plotly_fig(df1,'name','editcount','editcount',count=False)
fig = analyzer.create_plotly_fig(df1,'name','deletecount','deletecount',count=False)
fig.write_image("./data/figure_3.png")

ToDo

GroupChat Analysis

other task yet to accomplish are :

basicStats : It will return some basic stats of group. Such as total users, total messages, total media messages and total link shared.
wordCloud : It will create a word cloud, through which we can easily understand the most frequent words used in chat.
mostActiveUsers : It will create a bar chart for the top-10 most active members in the group.
mostActiveDay : It will create a bar chart to show traffic on whatsapp group at each weekdays.
topMediaContributor : It will create bar chart to show top-10 media contributers in the group.
maxWordContributers : It will create bar chart which will show top-10 authors who used max no. of words in their messages.
maxURLContributers : It will create bar chart to show top-10 url contributers in the group.
mostActiveTime : It will create bar chart to show the time at which group was highly active.
mostSuitableHours : It will create a bar chart to show the best time span at which there may be high chances getting responce from other group members.
wordCloud_in : It will create word cloud of particular individual. You need to just pass user name.
highlyActiveDates : It will create a bar chart to show highly active top-15 dates.
timeseriesAnalysis : It will plot user interactive time-series plot on traffic at each day.
activeMonthsB : It will create a bar chart to show most active months on which group was highly active.
maxEmojiUsers : This will plot a bar chart to show top-15 users who used max no. of emojis in group.
trafficPerYear : It will plot a bar chart to show traffic on group per year.
activeMonthsT : This will create a timeseries plot to show traffic on group per month.
weekdaysTraffic : It will create heat map to show the weekdays traffic along with time span.
topEmojis_G : It will return pandas dataframe of top-20 emojis used by users.
topEmojis_I : It will return top-10 emjois used by individual user.
saveDatframe : It will save the preprocessed data as csv file.

Using plotnine.ggplot

import numpy as np
import pandas as pd
from plotnine import *

%matplotlib inline
#df1
#ggplot(df1, aes(x='month', y='mlen')) + geom_point()

import pandas as pd 
from plotnine import ggplot, aes, facet_grid, labs, geom_col 
  
# reading dataset 
xd, yd, color = "day","mlen", 'day'
#+ facet_grid(facets="~year")
#p1.save(filename="./data/figure_1.png", quiet=True)
#plot = ggplot(df1, aes(x=xd, y=yd, fill=color)) + labs( x=xd, y=yd) + geom_col()
#plot = ggplot(df1, aes(x='date_time')) + geom_line(stat='count', color='blue') 
#plot = ggplot(df1, aes(x='dow', y='mlen')) + geom_line(stat='sum', color='blue') 
#plot = ggplot(df1, aes(x='dow', y='mlen')) + geom_bar(stat='sum')
#plot = ggplot(df1, aes(x='month', y='mlen',color='mlen')) + geom_point() 
#plot = ggplot(df1, aes(x='emojicount')) + geom_histogram(binwidth=1, fill='blue', color='black', alpha=0.7) 
#plot = ggplot(df1, aes(x='monthn', y='editcount')) + geom_boxplot()
#plot = ggplot(df1, aes(x='dow', y='hour')) + geom_tile(aes(fill='mlen'), color='white') + scale_fill_gradient(low='white', high='blue') 
#plot = ggplot(df1, aes(x='day')) + geom_bar(stat='count', fill='green') + coord_flip()
#plot = ggplot(df1, aes(x='month')) +geom_bar(stat='count', fill='green') + coord_flip()
#plot = ggplot(df1, aes(x='mlen')) +geom_histogram(binwidth=10, fill='orange', color='black', alpha=0.7) 

#plot = (ggplot(df1, aes(x="hour", y="emojicount", color="ym")) +geom_point() +facet_wrap("~year") +labs(x="Hour of Day", y="Emoji Count", title="Emojis by Hour and Month"))
plot = (ggplot(df1, aes(x="hour", y="emojicount", color="ym")) +geom_point() +facet_wrap('~ym') +labs(x="Hour of Day", y="Emoji Count", title="Emojis by Hour and Month"))
plot = (ggplot(df1, aes(x="mlen")) + geom_histogram(bins=30) +labs(x="Message Length", y="Count", title="Distribution of Message Lengths"))
plot = (
    ggplot(df1, aes(x="mlen", color="dow")) +
    geom_histogram(bins=30) +
    facet_wrap('~dow') +
    labs(x="Message Length", y="Count", title="Message Length Distribution by Name")
)

print(plot)

from plotnine import ggplot, geom_point, aes, stat_smooth, facet_wrap
from plotnine.data import mtcars

(ggplot(mtcars, aes("wt", "mpg", color="factor(gear)"))
 + geom_point()
 + stat_smooth(method="lm")
 + facet_wrap("~gear"))