International Journal for Research in Engineering Application & Management (IJREAM)

ISSN : 2454-9150 Vol-10, Issue-01, Apr 2024

Whatsapp Chat Analysis Based On NLP Using

ML
1
Dr.Sneha Jondhle, 2Mr.Harsh Jondhle,3Miss.Insha Mulla , 4Mr.Ibad Nachan 5Mr.Prashik Tayde
1
Associate Prof, 2Asst.Prof,3,4,5UG Student,1,2EXTC Engg, 3,4,5Computer Engg.
1
Shivajirao S. Jondhale College of Engineering, Dombivli, 2,3,4,5Shivajirao S. Jondhle
College of Engineering & Technology, Asangaon, Maharashtra, India.
1
[email protected],[email protected],[email protected],
4
[email protected], [email protected]
Abstract - WhatsApp has an extremely user-friendly user interface, which is one of its main features. WhatsApp is
the first messaging app that springs to mind when you want to connect or interact with someone. Everyone finds it
convenient to communicate via WhatsApp. To send a message, just open the app, choose Contact, and start typing.
You're done there. The user can utilize WhatsApp chats to examine discussions with peers, customers, or big groups
of people, as the messaging app receives over 65 billion messages daily. To discover fascinating details about the
most active people, the ghosts (those who do conversation research on a college student WhatsApp group
conversation is used to uncover fascinating information like who is most active, who are ghosts (those who don't
reply often), the most popular emoji, the busiest times of the day, or the most common phrases. These would surely
be some enlightening findings from the chat data. [1]

Keywords- WhatsApp Chat, Python, Streamlit, Analysis, Nature Language Processing, Emoji, Pandas, Matplotlib

I. INTRODUCTION II.AIMS AND OBJECTIVE

Web applications can be developed on a server using a) Aim

Python. It can be used in conjunction with programs to
The project makes sure to offer a thorough exploratory data
create workflows.. Python allows for execution of intricate
analysis on a range of WhatsApp chats
mathematical operations. Numerous Python libraries are
utilized, such as: An open-source Python framework called b) Objective
Stream-lit that enables users to create and distribute
Recognizing consumer behavior: Businesses can learn
visually appealing, complex data apps may be developed
about the requirements, preferences, and behavior of their
and implemented.
customers by examining chat interactions. This data
The main purpose of the open-source pandas library is to
can be utilized to enhance product offers, create more
make dealing with relational data efficient. It offers a range
focused marketing campaigns, and provide better customer
of methods and data structures for handling numerical and
service.
time series data. Based on NumPy arrays,A multiplatform
data visualization tool is called Matplotlib ,toolkit intended Enhancing team collaboration: Teams can better understand
to be used with the extensive SciPy stack. A Python class communication patterns, pinpoint bottlenecks, and enhance
called URL Extract gathers (extracts) URLs from input. communication by working on a chat analysis project.
Emoji analysis makes use of emoji. WhatsApp chat review
Better results and increased productivity may
is also great important application as it helps a lot as most
result from this and help the one to reach the correct taught
of the time, a person buys or cancel a product only based on
which increases productivity.
reviews. Thus, it is clear to show the importance of reviews.
Although, it will be quite difficult to go through thousands III.LITERATURE SURVEY
of reviews whenever a person thinks of buying a product.
Paper 1: “Content Analysis of WhatsApp
Thus, it will be good to scratch out useful info from these
Conversations: An Analytical Study to Evaluate the
reviews [1][10]
Effectiveness of WhatsApp Application in Karachi”.
As to the August 4, 2016, publication of the UK
Communications Market Report (CMR), instant messaging

26 | IJREAMV10SSJ2406 DOI : 10.35291/2454-9150.2024.0256 © 2024, IJREAM All Rights Reserved.

 International Journal for Research in Engineering Application & Management (IJREAM)
ISSN : 2454-9150 Vol-10, Issue-01, Apr 2024

has emerged as a highly favored mode of communication. Paper 3: “WhatsApp Group Data Analysis with R.”
In 2016, the percentage of individuals utilizing instant
The WhatsApp group chat dataset used for the analysis
messaging services increased from 28% to 43%. In the UK,
spans a year, from May 2015 to May 2016, with a total of
there is a fall in texting and emailing for the first time.
5,5563 records. It includes various attributes that indicate
According to 82% of respondents, using smartphone apps
the extent to which a specific user uses the group, including
like SnapChat and WhatsApp facilitates happier and
the number of years of usage, the amount of time spent
simpler communication with family and friends. As a
using it each day, response levels, the types of messages
result, the number of users of instant messaging apps has
posted by group members (Smiley, Text, and Count) which
significantly increased. According to the study's findings,
age-groups have the most active members, and so forth.The
Facebook Messenger has the most user base (22.5 million),
primary attributes used in this analysis are th age sent, how
followed by WhatsApp (16.7 million).
long it has been used for in a year, month, week, day, or
An investigation into the potential impact of instant hour, the timestamp (AM/PM), the senders' age group, and
messaging on interpersonal relationships was conducted their gender (male or female).Because it is open-source,
qualitatively at the Chinese University of Hong Kong. In RStudio is the most widely used integrated development
this context, ten informants of various ages and environment for R and is used for exploratory data analysis
backgrounds participated in a structured interview. The and visualization of the gathered data. [5]
interviewees listed a number of advantages and
IV.EXISTING SYSTEM
disadvantages of instant messaging. Several informants
highlighted the benefits of affordaability, effectiveness, The existing system has undergone significant
portability, and simplicity of use.[3] development. No features for sharing documents, sharing
locations, or displaying status were available in the
Paper 2: “Survey Analysis on the usage and Impact of
previous version. All of these features are present in the
WhatsApp Messenger”
current version.
These studies cover WhatsApp's effects on students and
The user were unable to exchange photographs via the Doc
young people. According to the survey, people in the range
format on previous versions. This approach allows users to
of 18 and 23 in the southern region of India use WhatsApp
use WhatsApp on Windows via the web application, which
for roughly 8 hours per day and are occasionally online for
can be connected via QR code. Another function is export
up to 12 to 16 hours per day. The majority of them
chat, which allows users to communicate, share, or get chat
acknowledged that WhatsApp was their preferred website.
information for analysis of data. A The latest version also
They share videos, audio, and photos. This study also
introduces a robust export chat feature, enabling to analyze
demonstrated that WhatsApp is the program that is used on
chat data across various platformsPeople rely on online
smartphones the most, compared to all other apps. Since
products from food to cloth and from home to electronics,
WhatsApp is the most popular app among young people
rather than going outside. Several products are available on
and other generations, this project can give them insights
these platforms by different brands. Thus, it will be quite
into their discussions and reveal unknown truths to
difficult to choose a product that is useful and reliable. To
them.This study was done to determine WhatsApp usage's
get a useful product, a user goes through the reviews of the
benefits and drawbacks. As this survey has revealed, of
product, to understand the product and to decide whether to
course. [4]
purchase it or not. [2][10]
V.COMPARATIVE STUDY
Sr. System Title Authors Objective Methodology /Technique Result/Performance
No.

1. Analyzing WhatsApp Vanshika Kooper, Analyse WhatsApp chats Semantic Analysis For Analysing WhatsApp chat data
Chat Using Python Bhoomi Gupta, using Python libraries to
Libraries Deepika Bansal. derive insights into user
behaviour and
communication patterns.

2. Content Analysis of Analyze WhatsApp usage: Matplotlib and Seaborn, WhatsApp is integral for students
WhatsApp Conversations: Sana Shahid message frequency, media Word Cloud, and professionals, facilitating
An Analytical Study to sharing, day-night activity, Word Cloud Generation, communication
Evaluate the Effectiveness conversation differences. Sentiment Analysis
of WhatsApp Application
in Karachi
3. Survey Analysis on the Naveen Kumar Investigate the usage Pandas, NumPy, . Limited Indian literature on
usage and Impact of and Sudhansh patterns and impact of Communication tools, Data WhatsApp usage, but studies
WhatsApp Messenger Sharma WhatsApp Messenger analysis tools. indicate increased connectivity and

27 | IJREAMV10SSJ2406 DOI : 10.35291/2454-9150.2024.0256 © 2024, IJREAM All Rights Reserved.

 International Journal for Research in Engineering Application & Management (IJREAM)
ISSN : 2454-9150 Vol-10, Issue-01, Apr 2024

through survey analysis. intimacy through the app,

particularly among youth
4. WhatsApp Group Data Sanchita Patil Project aims to understand Methodology involved exploratory Insights included user activity
Analysis with R user behavior and data analysis and visualization in trends, response levels, and
engagement patterns. RStudio demographics.

VI.PROBLEM STATEMENT reporting, the proposed system would be able to provide

more accurate and thorough knowledge about user
WhatsApp-Analyzer provides statistical analysis for
behavior and preferences. putting data security and privacy
WhatsApp talks. WhatsApp conversation files can be
first , the proposed approach will also boost user
exported to make charts indicating which participants a user
confidence and ensure that the analysis is conducted in an
most frequently interacts with. This project uses dataset
ethical and responsible manner. The proposed
modification approaches to improve understanding
methodology holds significant potential to enhance the
Whatsapp Communication on our phones.[1].
utility and worth of WhatsApp Chat Analysis for
VII. PROPOSED SYSTEM researchers and enterprises.
The suggested solution would use automatic data VIII. ALGORITHM
extraction techniques to compile information from
1. User Interface Setup:
WhatsApp talks. This would mean using specialized
Display the Streamlit sidebar with the title "WhatsApp
software or algorithms to extract the data straight from
Chat Analyzer".
WhatsApp servers, rather than requiring users to copy and
Provide a file uploader for users to upload a WhatsApp
paste chat logs by hand. This approach would reduce the
chat data file.
likelihood of biases or errors and ensure that the data is
import streamlit as st
collected consistently and reliably. After the data of
st.sidebar.title("WhatsApp Chat Analyzer")
WhatsApp chat has been retrieved, the proposed system
uploaded_file= st.sidebar.file_uploader("Choose a file")
will apply advanced data cleaning and pre-processing
techniques to get it ready for analysis.In order to prepare 2. Data Processing:
the data for analysis, it would be necessary to eliminate Read the uploaded chat data.
duplicate data, deal with missing values, and arrange the Preprocess the data to extract useful information like user,
data in a systematic manner. The suggested system would message content, date, etc.
use state-of-the-art data cleaning and pre-processing import preprocessor
techniques to confirm that the data is of the highest quality if uploaded_file is not None:
and ready for more in-depth analysis. It would use bytes_data = uploaded_file.getvalue()
advanced analysis techniques to mine the data of data = bytes_data.decode("utf-8")
WhatsApp chat to identify patterns and insights. This df=
would mean using methods for natural language processing preprocessor.preprocess(data)cnn_model.add(Dense(128,
(NLP) to identify keywords and the conversation's activation='relu'))
emotional undertone in addition to machine learning cnn_model.add(Dense(2, activation='softmax'))
algoritnms to identify trends and patterns in the data.By
3. User Selection and Analysis Trigger:
utilizing these techniques,the suggested system would be
Allow the user to select a specific user for analysis.
able to provide more accurate and comprehensive insights
Trigger the analysis upon clicking the "Show Analysis"
regarding user behavior and preferences. sophisticated
button.
analytic methods on pages. Thanks to interactive
if st.sidebar.button("Show Analysis"):
visualization and reporting capabilities, users may review
# Perform analysis...
the data and see the outcomes of the research. This would
require the creation of interactive dashboards and reports 4. Analysis and Visualization
that enables users to investigate more deeply and examine Calculate statistics like total messages, words, shared
the data in more depth. By providing more dynamic and media, and links shared.
user-friendly visualization and reporting features, the Display the statistics using Streamlit components.
suggested system would aid users in understanding and import helper
interpreting the study's findings. Several essential elements num_messages, words, num_media_messages, num_links
of the proposed WhatsApp Chat Analysis system = helper.fetch_stats(selected_user, df)
contribute to the process's increased scalability, accuracy, st.title("Top Statistics")
and speed.By leveraging automated data extraction, st.header("Total Messages")
advanced analysis methods, interactive visualization, and st.title(num_messages)

28 | IJREAMV10SSJ2406 DOI : 10.35291/2454-9150.2024.0256 © 2024, IJREAM All Rights Reserved.

 International Journal for Research in Engineering Application & Management (IJREAM)
ISSN : 2454-9150 Vol-10, Issue-01, Apr 2024

# Display other statistics... Media Sharing:

(M = \text{count_media_files})
5. Data Processing and Analysis:
Busiest Days and Months:
Import necessary libraries and modules.
(T_d): Plot daily chat activity.
Define functions for statistical analysis, semantic analysis,
(T_m): Plot monthly chat activity.
etc.
User Analysis:Most active user:
from urlextract import URLExtract
(max(A_i))Interactionmatrix:C{ij}
from wordcloud import WordCloud
from collections import Counter X.SYSTEM ARCHITECTURE
import pandas as pd
import re
import emoji
import matplotlib.pyplot as plt
import seaborn as sns
from model import load_sematic_analysis_pipeline
from model import load_sentiment_analysis_pipeline
import numpy as np
from sklearn.cluster import KMeans
import streamlit as st
6. Semantic Analysis:
Define functions for semantic analysis using pre-trained
models. Fig.1: System Architecture
Perform clustering and sentiment analysis on the chat
data. Here's a step-by-step guide for the architecture flow in the
def plot_sematic_analysis(selected_user, df): WhatsApp Chat Analysis Project:
# Semantic analysis implementation... 1. Data Extraction:
Obtain chat data directly from WhatsApp servers through
7. Streamlit Integration: human or automated means.
Set up Streamlit interface with sidebar and file uploader. Store the extracted data securely in a database for further
Perform data processing and analysis upon file upload. processing.
Integrate semantic analysis into the Streamlit app. 2. Data Pre-processing and Cleaning: Organize and
if __name__ == "__main__": transform the data for analysis.
st.sidebar.title("WhatsAppChat Analyzer") Clean the data by resolving missing values and eliminating
uploaded_file= st.sidebar.file_uploader("Choose a file") noise, and structuring the data appropriately.
# Display statistics... 3.Natural Language Processing (NLP):
# Perform semantic analysis Analyze the text data from WhatsApp chats using NLP
plot_sematic_analysis(selected_user, df) techniques.
IX.MATHEMATICAL MODEL Identify keywords, gauge sentiment, and recognize specific
Let’s denote the following variables: entities mentioned in the chats.
(N): Total messages exchanged in the chat. 4. Machine Learning:
(W): Total word count across all messages. Apply machine learning methods to discover patterns and
(M): Number of media files (images, videos, etc.) trends within the data.
shared. For labeling, use classification algorithms and clustering
(T_d): Timeline of chat activity (daily basis). algorithms to group related discussions.
(T_m): Timeline of chat activity (monthly basis). Create predictive models using past data to forecast future
(U_i): User (i) (where (i = 1, 2, \ldots, n)) in the chat. trends.
(A_i): Count of messages sent by user (i). 5. Visualization and Reporting:
(C_{ij}): Count of messages from user (i) to user (j). Create interactive dashboards and reports to present
(W_f): Word frequency distribution. analysis results.
(E): Emoji analysis. Use visualization equipments are used to generate charts
Now, let’s express some key insights: and graphs.
Total Chat Activity: 6. Data Security and Privacy:
(N = \sum_{i=1}^{n} A_i) Implement robust Limits on access and encryption to
Word Count: safeguard information from unwanted access.
(W = \sum_{i=1}^{n} Adhere to industry best practices for data privacy and
\text{word_count}(U_i)) security.

29 | IJREAMV10SSJ2406 DOI : 10.35291/2454-9150.2024.0256 © 2024, IJREAM All Rights Reserved.

 International Journal for Research in Engineering Application & Management (IJREAM)
ISSN : 2454-9150 Vol-10, Issue-01, Apr 2024

This structured approach enables businesses and

researchers to gain applicable insights from WhatsApp
chat data while upholding high standards of data security
and privacy.
XI.ADVANTAGES
 Sentiment analysis drives proactive feedback
response and boosts customer and employee
satisfaction. Fig 4: Top Statistics

 Automated responses and issue prioritization

streamline customer support, enhancing response
times and resolution efficiency.
 Chat data analysis informs trend identification and
market research, facilitating informed decision-
making and adaptive strategies.
 Feedback extraction and anomaly detection aid in
product improvement and fraud detection,
Fig 5: Activity Map
respectively, fostering trust and product integrity.
XIII.CONCLUSION
 Multilingual support through NLP-driven
translation expands market reach and fosters Thus this research try to impement the paper “Analyzing
global communication, positioning businesses for WhatsApp Chat Using Python Libraries” shah, Amar
sustained growth and inclusivity. Jaiswal.IEEE,2023 it has concluded that data analysis and
visualization of the exported WhatsApp chat may be
XII.DESIGN DETAILS accomplished using Python and its libraries. Whatsapp is a
popular tool for communication. As a outcome, it contains a
large amount of chat data that may be examined to learn
more. Only members of WhatsApp groups may collect chat
data, and it is presumed that the sender has not removed
any messages to avoid misleading results from the research.
The app displays the intended outcome,
such as the monthly A range of data can be visualized with
the matplotlib package, including timeliness, daily
timeliness, busiest month and day, most active user, weekly
activity map, etc. The user can also be able to analysis the
word which he had used most of the time in the
Fig 2: Wordcloud
conversation as well as the user will able to analyse the
emoji and analyse the behaviour of the chat. There may be
additional room in the chat analysis for one to look for
other pertinent information.
XIV.REFERENCE
[1] Rajan,Bhoomi Gupta,Vanshika Kapoor,Deepka
Bansal Department of IT Maharaja Agrasen
Institute of Technology “Analyzing WhatsApp
Chat Using Python Libraries” shah, Amar
Jaiswal.IEEE,2023
Fig 3: Emoji Analysis [2] Ravishankara K, Dhanush, Vaisakh, Srajan I S,
“International Journal of Engineering Research &
Technology (IJERT)”, ISSN: 2278-0181, Vol. 9
Issue 05, May-2020
[3] Sana Shahid. “Content Analysis of Whatsapp
Conversations: An Analytical Study to Evaluate
the Effectiveness of Whatsapp Application in

30 | IJREAMV10SSJ2406 DOI : 10.35291/2454-9150.2024.0256 © 2024, IJREAM All Rights Reserved.

 International Journal for Research in Engineering Application & Management (IJREAM)
ISSN : 2454-9150 Vol-10, Issue-01, Apr 2024

Karachi". International Journal of Media,

Journalism and Mass Communications (IJMJMC),
vol 4, no. 1, 2018, pp. 14-26.
doi:http://dx.doi.org/10. 20431/2455-
0043.0401002
[4] Naveen Kumar and Sudhansh Sharma “Survey
Analysis on the Usage and Impact of Whatsapp
Messenger”, Global Journal of Enterprise
Information System. Volume-8, Issue-3,July-
September,2016.
(http://informaticsjournals.com/index.php/gjeis)
[5] Sanchita Patil. “Content Analysis of Whatsapp
Conversations: WhatsApp Group Data Analysis
with R". International Journal of Computer
Applications (0975 – 8887) Volume 154 – No.4,
November 2016.
[6] Ahmed, I., Fiaz, T., “Mobile phone to youngsters:
Necessity or addiction”, African Journal of
Business Management Vol.5 (32), pp. 12512-
12519, Aijaz, K. (2011).
[7] Aharony, N., T., G., The Importance of the
WhatsApp Family Group: An Exploratory
Analysis. “Aslib Journal of Information
Management, Vol. 68, Issue 2, pp.1-37” (2016).
[8] F. Meng Cai, “PubMed Central”, PMCID:
PMC7944036, PMID: 33732917.
[9] Radhika, Narendiran, “Kind of Crops and Small
Plants Prediction using IoT with Machine
Learning,” International Journal of Computer &
Mathematical Sciences April 2018, pp. 93-97.
[10] Prof. Satish Manje, “Sentiment Detection for
Amazon Product Review” in IJREAM, ISSN :
2454-9150, Volume09,Issue 01,Special Issue,
April 2023.

31 | IJREAMV10SSJ2406 DOI : 10.35291/2454-9150.2024.0256 © 2024, IJREAM All Rights Reserved.