Social media analysis using vader sentiment tool

Project Phase 1: Progress 1-Presentation
on
“Social Media Analyzer”
STUDENTS :
Sanath Kumar N (1RN20CS128)
Sanket Krishna Hegde (1RN20CS131)
Shreyas BR(1RN20CS138)
Guide name: Devraju BM
Prof./Asso Prof/Asst. Prof. Dept. of CSE,
RNSIT
Department of Computer Science and Engineering
RNS Institute of Technology
2023-24

CONTENTS
• INTRODUCTION
• REQUIREMENTS
• SYSTEM ARCHITECTURE/BLOCK DIAGRAM
• IMPLEMENTATION
• RESULTS & CONCLUSION
• REFERENCES
Dept. of CSE,RNSIT 2023 - 24 2

INTRODUCTION
• LITERATURE SURVEY
• PROBLEM STATEMENT OF PROPOSED PROJECT
• EXISTING SYSTEM AND THEIR LIMITATIONS
• OBJECTIVES OF PROJECT WORK.

LITERATURE SURVEY
4
Dept. of CSE,RNSIT 2023 - 24

PROBLEM STATEMENT OF PROPOSED PROJECT
• The need arises for an efficient and accurate sentiment analysis tool specifically tailored for
Facebook comments. This tool should seamlessly integrate with Facebook's data structure,
providing real-time or near-real-time analysis to capture dynamic sentiment changes. Existing
sentiment analysis tools may lack the capability to fully understand the intensity and polarity of
sentiments expressed in user comments, limiting their effectiveness in providing nuanced insights.
Moreover, challenges such as handling slang, adapting to evolving language trends, and ensuring
scalability further contribute to the limitations of current systems.
• The proposed project aims to address these challenges by developing a robust sentiment analysis
system using the VADER sentiment analysis tool. The system will not only automate sentiment
classification but also provide comprehensive user Interface for view account and Post statistics
related to the user. Integration with Facebook's data structure will enable the analysis of large
volumes of comments, allowing users to gain valuable insights into the sentiment distribution,
comment intensity trends, and overall sentiment patterns associated with specific accounts or posts.
5

EXISTING SYSTEM AND THEIR LIMITATIONS
• 1. Accuracy and Contextual Understanding:
• Many existing sentiment analysis systems may struggle with accurately understanding the context of user-
generated content, leading to misinterpretations of sentiment.
• 2. Scalability Challenges:
• Traditional sentiment analysis methods may face challenges in scaling to handle the vast amount of data
generated on social media platforms like Facebook, resulting in slower processing times.
•
3. Inability to Capture Intensity:
• Some sentiment analysis tools might provide binary sentiment labels (positive/negative) but may lack the
capability to capture the intensity or strength of the sentiment expressed in a comment.
• 4. Limited Adaptability to Slang and Emojis:
• Existing systems may not be well-equipped to handle the use of slang, colloquial expressions, or emojis
commonly found in social media language, potentially leading to misinterpretation
• 5. Dependency on Predefined Lexicons:
• Systems relying on predefined sentiment lexicons may struggle with staying updated and may not effectively
adapt to evolving language trends and the emergence of new words.
6
2023 - 24
Dept. of CSE,RNSIT

7
OBJECTIVES OF PROJECT WORK
1. Automated Sentiment Analysis:
•Develop an automated sentiment analysis system using the VADER sentiment analysis tool for Facebook comments.
•Achieve accurate sentiment classification for diverse user-generated content on the platform.
2. Integration with Facebook:
•Seamlessly integrate the sentiment analysis tool with Facebook's data structure and comment formats.
•Ensure real-time or near-real-time analysis to capture dynamic changes in sentiment.
3. Comprehensive Sentiment Scores:
•Implement the polarity_scores() method to obtain comprehensive sentiment scores, including positive, negative, neutral, and
compound scores for each comment.
•Provide detailed insights into the sentiment distribution of Facebook comments.
4. Statistical Analysis and Reporting:
•Generate comprehensive reports for each analyzed account or post, including statistical metrics such as mean sentiment,
sentiment distribution, and comment intensity trends.
•Provide visualizations and graphical representations for easier interpretation of sentiment patterns.
5. User-Friendly Interface:
•Design a user-friendly interface for interacting with the sentiment analysis tool.
•Ensure easy navigation and accessibility for users to input data and retrieve sentiment analysis results.
6. Scalability and Performance:
•Build the system to handle large-scale data efficiently, ensuring scalability for accounts or posts with a high volume of comments.
•Optimize performance to deliver timely results without compromising accuracy.
7. Documentation and Support:
•Create comprehensive documentation for users, detailing how to use the sentiment analysis tool effectively.
•Provide user support to address any queries or issues encountered during the analysis process.

REQUIREMENTS
• Hardware Requirements
• Software Requirements

HARDWARE REQUIREMENTS
● Processor : Intel Duo Core
● RAM : 8GB
● GPU : NVIDIA Tesla K80
9
2023 - 24
Dept. of CSE,RNSIT

SOFTWARE REQUIREMENTS
• Operating System : Ubuntu (14) (16 bit) , Windows 10
• Programming Languages : Python , Java Script ,React , nltk.
• ML Framework : Vader .
• Server : FaceBook,Heroku
10
2023 - 24
Dept. of CSE,RNSIT

REQUIREMENTS SPECIFICATION:
1. React.js for front-end development
2. Socket.IO for real-time communications between clients
3. Node.js for server-side development
4. Express.js for building server-side APIs
5. MongoDB for storing user data and call records
6. IDE or text editor of your choice (Visual Studio Code, Sublime Text, etc.)
7. Optional: CSS framework like Bootstrap or Material UI for styling 11.
8. Backend technologies for data processing
9. Machine learning libraries for sentiment analysis
10. APIs for accessing social media data.
11

SYSTEM ARCHITECTURE/BLOCK
DIAGRAM
• Block diagram/Architecture diagram/DFD/Sequence diagram/Flowchart

BLOCK DIAGRAM
13

ARCHITECTURE DIAGRAM
14

DATA FLOW DIAGRAM
15

USEFULNESS OF THE PROJECT
Dept. of CSE,RNSIT 2023 – 24 16

CONTINUED …..
17

IMPLEMENTATION DETAILS
Pseudo-code implementation of the first half:
NLTK:is used for understanding of human natural language.
Matplotlib: is a python module used for data visualization and and 2D plotting for
representation of data.
NumPy: is a python package used for scientific and computional methods in python.
Pandas: is a python module used for data preprocessing and analysis .
->pip install numpy,matplotlib,pandas,nltk

19
Pseudo-code
1) Import necessary Libraries
import time
import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import nltk
import io
import unicodedata
import numpy as np
import re
import string
from numpy import linalg
import socket
socket.getaddrinfo('localhost', 8080)
from nltk.sentiment.vader importSentimentIntensityAnalyzer
from nltk.tokenize import sent_tokenize, word_tokenize
from nltk.tokenize import PunktSentenceTokenizer
from nltk.tokenize import PunktSentenceTokenizer
from nltk.corpus import webtext
from nltk.stem.porter import PorterStemmer
from nltk.stem.wordnet import WordNetLemmatizer
nltk.download('punkt')
nltk.download('wordnet')
nltk.download('averaged_perceptron_tagger')
nltk.download('vader_lexicon')

20
2) Youtube-API setup
import googleapiclient.discovery
import googleapiclient.errors
api_service_name = "youtube"
api_version = "v3"
DEVELOPER_KEY = "AIzaSyBWZoWKKtgc5s-oHTE5P_l0FwafRENBkow"
VIDEO_ID = "-1gF6PIB1YQ" # <<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<<< Enterthe video id here
youtube = googleapiclient.discovery.build(
api_service_name, api_version, developerKey=DEVELOPER_KEY)
request =youtube.commentThreads().list(
part="snippet",
videoId=VIDEO_ID,
maxResults=100
)
response = request.execute()
text = ''
for item in response['items']:
# print(item['snippet']['topLevelComment']['snippet']['textDisplay'])
text += item['snippet']['topLevelComment']['snippet']['textDisplay']

21
3) After we open a file we preprocess the text through tokenize, stemize and then lemmatize:
Tokenize the text, i.e split words from text.
sent_tokenizer = PunktSentenceTokenizer(text)
sents = sent_tokenizer.tokenize(text)
print(word_tokenize(text))
print(sent_tokenize(text))
4)Stemize and lemmatize the text for normalization of the text:
a)For stemize we use PorterStemmer() function:
from nltk.stem.porter import PorterStemmer
porter_stemmer =PorterStemmer()
nltk_tokens = nltk.word_tokenize(text)
for w in nltk_tokens:
print (“Actual: %s Stem: %s” % (w, porter_stemmer.stem(w)))
b) For lemmatize we use WordNetLemmatizer() function :
from nltk.stem.wordnet import WordNetLemmatizer
wordnet_lemmatizer = WordNetLemmatizer()
nltk_tokens = nltk.word_tokenize(text)
for w in nltk_tokens:
print (“Actual: %s Lemma:%s” % (w, wordnet_lemmatizer.lemmatize(w)
Actual: This Stem: thi
Actual: is Stem: is
Actual: very Stem: veri
Actual: Bad Stem: bad
Actual: and Stem: and
Actual: good Stem: good
Actual: . Stem: .
Actual: This Lemma: This
Actual: is Lemma: is
Actual: very Lemma: very
Actual: Bad Lemma: Bad
Actual: and Lemma: and
Actual: good Lemma: good
Actual: . Lemma: .
['This', 'is', 'very', 'Bad', 'and', 'good', '.']
['This is very Bad and good.’]

22
5) POS( part of speech) tagging of the tokens and select only significant features/tokens like
adjectives, adverbs, and verbs, etc.
text = nltk.word_tokenize(text)
print(nltk.pos_tag(text))
6)Pass the tokens to a sentiment intensity analyzer which classifies the Facebook comments as
positive, negative or neutral.
sid = SentimentIntensityAnalyzer()
tokenizer = nltk.data.load(‘tokenizers/punkt/english.pickle’)
with open(‘kindle.txt’,encoding=’ISO-8859-2′) as f:
for text in f.read().split(‘n’):
print(text)
scores = sid.polarity_scores(text)
for key in sorted(scores):
print(‘{0}: {1}, ‘.format(key, scores[key]), end=”)
print()
This is very Bad and good.
compound: -0.1604, neg: 0.346, neu: 0.365, pos: 0.289,
[('This', 'DT'), ('is', 'VBZ'), ('very', 'RB'), ('Bad', 'NNP'), ('and', 'CC'), ('good', 'JJ'), ('.', '.')]

RESULTS & CONCLUSION
• Screenshots of Results-

24
Limitations of Project work implemented :-
1.Static Video ID:
The script uses a static VIDEO_ID variable to fetch comments for a specific video. This limits the script's flexibility, as it would need
to be modified manually each time you want to analyze comments for a different video.
Hardcoded API Key:
2.The YouTube Data API key (DEVELOPER_KEY) is hardcoded in the script. This could pose a security risk if the code is shared or
made publicly accessible, as the API key grants access to the associated YouTube account's data. Best practices involve securing
API keys and not hardcoding them directly in the code.
3.Rate Limiting and Quotas:
The YouTube Data API has usage quotas and rate limits. If the script makes too many requests in a short period, it may hit these
limits and result in temporary restrictions. Proper rate-limiting strategies and error handling for quota issues are essential for a
robust implementation.
Data Processing:
4.The script collects comment text but does not perform any meaningful analysis on the sentiment or other aspects of the
comments. Depending on the project's goals, additional processing steps may be necessary to derive insights from the collected
data.
5.project involves retrieving comments from a Facebook page, the access token handling in the script may need further
consideration. Facebook Graph API access tokens have expiration periods, and refreshing them programmatically may be required.
Code Modularity:
6.The script lacks modularity, making it less reusable for different scenarios. Encapsulating functionalities into functions or classes
could enhance code organization and maintainability..

REFERENCES/BIBLIOGRAPHY
[1]

SUGGESTIONS….!

THANK YOU!!!

Social media analysis using vader sentiment tool

More Related Content

Similar to Social media analysis using vader sentiment tool

Recently uploaded

Social media analysis using vader sentiment tool