Content uploaded by Shobana Govindasamy
Author content
All content in this area was uploaded by Shobana Govindasamy on Jun 03, 2019
Content may be subject to copyright.
International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-7 Issue-4S, November 2018
343
Published By:
Blue Eyes Intelligence Engineering
& Sciences Publication
Retrieval Number: E1989017519
Twitter Sentimental Analysis
Shobana G, Vigneshwara B, Maniraj Sai A.
Abstract: In this current era, social media plays a important
role in data exchange, sharing their thoughts. Emotional Effect
of a person maintains an important role on their day to day life.
Sentiment Analysis is a procedureof analyzing the opinions and
polarity of thoughts of the person. Twitter is a main platform on
sharing the thought's, opinion and sentiments on different
occasions. Twitter Sentimental Analysis is method of analyzing
the emotions from tweets (message posted by user in twitter).
Tweets are helpful in extracting the Sentimental values from the
user. The data provide the Polarity indication like positive,
negative or unbiassed values. It is focused on the person’s tweets
and the hash tags for understanding the situations in each aspect
of the criteria. The paper is to analyse the famous person’s id’s
(@realdonaldtrump) or hash tags (#IPL2018) for understanding
the mindset of people in each situation when the person has
tweeted or has acted upon some incidents. The proposed system is
to analyze the sentiment of the people using python, twitter API,
Text Blob (Library for processing text). As the results it helps to
analysis the post with a better accuracy.
Keywords: (@realdonaldtrump), (#IPL2018), Text Blob
(Library for Processing Text).
I. INTRODUCTION
In the past years, the young generation people are moving
towards the social media like Google Plus, WhatsApp,
Facebook, Twitter, etc. The social media is also revolving
with those people to get them involved by making current
trending insights concepts that is trending within a second.
In the recent years, the people are exposing their social
related issues through several social media by comments,
reviews, posts, hashtags, emoji’s, etc. which was followed
by many people and those tweets become popular soon.
Moreover, the social media is also bringing tremendous
opportunity platform for businesses to connect with the
consumers so easily. People rest on mostly user produced
content like, comments, over online for making the decision.
Example: if anyone has to buy a product or make a decision,
they initiallysearch its reviews online, converse about it on
social media. The content that is displayed for that product
is mainly taken into the point as well as the discussion in the
social media is also noticed and these made the way to make
our business a success. To automate our analysis based on
the reviews or comments in the social media by the people,
for a sentimental analysis. Sentimental Analysis (SA) is
introduced to the world to tell us the information is correct
or wrong in each scenario using the social media tags. Thus,
we can know about how world or people are reacting to
every aspect currently going in the world.
Revised Version Manuscript Received on 25 November 2018.
Shobana G, Assistant Professor, Department of Computer Science and
Engineering, Kumaraguru College of Technology, Coimbatore (Tamil
Nadu), India.
Vigneshwara B, UG Scholars, Department of Computer Science and
Engineering, Kumaraguru College of Technology, Coimbatore (Tamil
Nadu), India.
Maniraj Sai A, UG Scholars, Department of Computer Science and
Engineering, Kumaraguru College of Technology, Coimbatore (Tamil
Nadu), India.
Fig 1.1 System Architecture
The system architecture consists of the components
asshown in the figure 1.1 such as Tweets extraction from
twitter, preprocessing of data, feature extraction, Training
set are defined for the given analysis. The training set is
obtained by predefined set of positive or negative tweets
which can be done using naive Bayes or support vector
machine (SVM) and output obtained is positive, negative
tweets. The Classifier will classify the tweets according the
training set and regulates the polarity of the tweet as the
output.
In this paper, we are going to analysis the microblog
called as Twitter, classify the “tweets” into positive,
negative and neutral sentiment. We explore the method for
building such data using Twitter hashtags (e.g., #best
feeling, #DonaldTrump, #love) to identify positive,
negative, and neutral tweets to use for training three-way
sentiment classifiers. Thus, these tweets and the hashtags are
must for analyzing the thinking level of individual people.
II. LITERATURE SURVEY
Sentiment analysis is the process of analysis of the text
from many levels. First level is document level [3],the
classification task determine the class of an objectbased on
its attributes (Turney, 2002; Pang and Lee, 2004), and after
that it can analysed at the sentence level[5] for classifying
the sentence based on the negative, positive and neutral
sentiments (Hu and Liu, 2004; Kim and Hovy, 2004) and
next level is the phrase level[4] for definingif an expression
is unbiassed or polar and then remove uncertainty of
Twitter Sentimental Analysis
344
Published By:
Blue Eyes Intelligence Engineering
& Sciences Publication
Retrieval Number: E1989017519
meaning fromthe polarity of the polar expressions
(Wilson et al., 2005; Agarwal et al., 2009. Bermingham and
Smeaton(2010) and Pak and Paroubek (2010). Go et al.
(2009) they used distant learning algorithm to obtain the
sentiment data [8]. In this techniques,positive emoticons
symbols in tweets such as “:)” “:-)”and negative emoticons
symbols in tweets such as like “:(” “:-(”. They proposed the
models using Naive Bayes algorithm for analysis the text
and the report are generated and visualized.
They used two methods such as unigrams for identifying
single word repeating over the context and bigrams for
identifying double word repeating over the context along
with Parts-of-Speech (POS) for analysing the tweets. But the
unigram method had reached a better way of analysis but the
bigrams and POS had failed to attempt his purpose.
Pak and Paroubek (2010) [2] collect the following tweets
considered as data which really helped them in similar
distant learning paradigm for setting a model for analysis.
They perform classification of task such as subjective,
objective. For subjective the informationare get from the
user tweets by means of text or image or symbols as Go et
al. (2009) [8]. For objective informationthe information are
obtained from verification of the data such as famous
newspapers like “Times of India”, “Washington Posts”
etc.Information which is taken for analysis is casual sample
of flowing tweets collected by using queries. In the past year
there have been numerousdocumentsobservingthe Twitter
sentiment and buzz [1], [2], [4] (Jansen et al. 2009; Pak and
Paroubek 2010; O’Connor et al. 2010; Tumasjan et al. 2010;
Bifet and Frank 2010; Barbosa and Feng 2010; Davidov,
Tsur, and Rappoport 2010). Furtherscientists have started to
discover the usage of part-of-speech structures but results
remain mixed. It has enormous interestingchances to
develop the innovative applications, because success of
many business depends on accessible information on online
sources such as blogs, twitter and other social networks.
Barbosa and Feng (2010)[4] has analysed the sentimental
classification on Twitter data.The test data of tweets are
collected, they have taken some of the syntax features for
analysis of tweets which contains symbols, retweet,
emoticons, tags, link, punctuation and exclamation marks,
semicolon are in the combination with structuresfor
identifying the polarity of words.
Kamps et al. (2002) [12] has analysed the data by using the
lexical database. Lexical database is description of lexemes.
Lexical database such as WordNet are used.This contains
the emotional content of a word. The distance metric of
words are used to determined semantic polarity of
adjectives.
Researchers are also trying to find different ways of
analysing tweets based on the ideas they had while
understanding the concept. Researchers tried this analysis
using some of the specified fields such as Machine learning
which uses Naive Bayes, Maximum entropy and SVM
alongside the Semantic Orientation based Word Net which
extracts equivalent words and similitude for the content
feature, then Lexicon based analysis based on the created
dataset which consists of pre-processed tweets and lastly,
Hybrid approach where some researchers combined the
supervised machine learning and lexicon based approaches
together to improve sentiment classification performance.
Gamon (2004) [9] has done sentiment analysis on feedback
data from the Global Support Services survey. They are used
query to identify the role of features like Part of Speech tags.
The accuracy of classifier can be obtained by some of
factors such as feature selection,from the testing data and
demonstrate the abstract linguistic analysis feature for
accuracy of data.
Devaki.p, et al (2017)[15] has done analysis on twitter data
for election. It indicates the popularity of parties in the
election based on positive tweets. This system uses Naïve
Bayes classifier algorithm are used to classify the positive
and negative tweets.
A comparative study of existing techniques for mining the
data which includes machine learning, Interdependent
Latent Dirichlet Allocation, lexicon-based approaches,
together with cross domain , cross-lingual methods and
some evaluation metrics. The concept level sentences
analysis uses the Combining Lexicon and Learning based
Approach. As the result of study, machine learning methods
such as Support Vector Machine and Naive Bayes have the
highest accuracy and can be regarded as the baseline
learning methods, while lexicon-based methods are very
effective in some cases.
More research is needed to determine whether the POS
features are just of poor quality due to the results of the
tagger or whether POS features are just less useful for
sentiment analysis in this domain. Features from an existing
sentiment lexicon were somewhat useful in conjunction with
microblogging features, but the microblogging features (i.e.,
the presence of intensifiers and positive/negative/neutral
emoticons and abbreviations) were clearly the most useful.
In this paper, we perform extensive feature analysis of
tweets using hashtags, ID’s and building model
classifications.
III. METHODOLOGY
In this method we uses textblob as a method to find the
polarity of the text ( positive text, negative text or neutral
text). The tweets are imported from the Twitter using the
(API) provided by the Twitter Developer. From these API
various fields like tweets, source, retweets, likes, language,
user etc. can be scrapped. After collecting these data, we can
analyses the various famous person thoughts on anevent or
occasion
Fig 3.1Architectural Flow of Twitter Analysis
International Journal of Recent Technology and Engineering (IJRTE)
ISSN: 2277-3878, Volume-7 Issue-4S, November 2018
345
Published By:
Blue Eyes Intelligence Engineering
& Sciences Publication
Retrieval Number: E1989017519
The figure 3.1explains the extraction of tweets id from
twitter through API, then preprocess the data that are
extracted. Preprocessing includes exclusion of unwanted
fields, segregating the fields important for analysis. Once the
fields are extracted and segregated CSV is created. Using
this CSV, the length of the message, Likes, retweets for the
id is extracted and various results are derived. With the
scraped tweets, classify the tweets whether positive or
negative or neutral.
A. Dataset Description
In this proposed system, we have used the dataset called
astwitterdataset. csv. It contains the following fields Tweets,
Len, ID, Date, Source, Likes, RT’s (Retweets), SA
(Sentimental Analysis).
B. Software Description
In the system the graphs such as Table, Bar graph,
Line graph are generated with the help of Spyder and
Jupyter notebook. The predefined functions are pandas,
numpy, matplotlib, pyplot, list, Dictionary. Pandas is used
for converting from csv file to dataset. Numpy is one of the
essential library for scientific calculating in Python. It
delivers a high-performance multidimensional array object,
and apparatuses for experementing with these arrays. Python
comprises ofnumerous built-in container categories: lists,
dictionaries, sets, and tuples. A list is the Python equal of an
array, but is resizable and can contain elements of different
types. A dictionary stores (key, value) pairs, like a Map in
Java or an object in JavaScript. Python library such as Text
Blobare used for processing the textual data. It provides API
for processing natural language processing (NLP) such as
part-of-speech tagging, noun phrase extraction, sentiment
analysis, classification, translation, and more. Tweepyisused
for accessing Twitter API and it is open sourced.
C. Data Analysis and Visualization
In Twitter the various famous personalities tweet their
thoughts on their opinion on an occasion. From their
thoughts, importance of that occasion and the polarity of
their tweet are analysed. Some of the analysis with the
dataset as follows.
Calculate the average length of the tweet and
visualize the average length for a period.
Visualize the favorites and retweets for each
personality.
Visualize the various source of the tweet.
Calculate the polarity of the tweets given by the
person
Visualize the Polarity of tweet (positive,
negative, neutral)
Compare the tweets polarity of various famous
personalities
Analyses to calculate the average length of the tweet
and visualize the average length for a period.
Calculate the average length of the tweet of the person
using the mean function and visualize the average length for
a period. With the calculated average. Visualize the average
length of tweet for a period.
Fig 3.2: Average length of Tweets for a given period
Analysistovisualize the favorites and retweets for each
personality.
For Visualizing the favorites and retweets for each
personality, it is needed to calculate the number of Favorites
and RT’s from CSV. Visualize the data and compare with
various personalities.
Fig 3.3: Narendra Modi Favorites and retweets
Fig 3.4 Rahul Gandhi Favorites and retweets
Analysis to visualize the various source of the tweet.
Twitter tweets are tweeted through various sources like
Media Source, iPhone or android phone. Visualize the
number of sources used by a user to tweet his opinion in
twitter. Most used source for tweets by the personalities can
be identified.
Fig 3.5. Source for Narendra Modi Tweets
Twitter Sentimental Analysis
346
Published By:
Blue Eyes Intelligence Engineering
& Sciences Publication
Retrieval Number: E1989017519
Fig 3.6 Sources for Rahul Gandhi Tweets
Analysis tocalculate the polarity of the tweets given by
the person
Using Textblob calculate the polarity of each tweets and
calculate the average polarity level of the tweets and
interfere that the person tweets polarity level. This allows to
find the polarity of the tweet on an occasion. Calculate and
store the three levels in the list.
Analysis tovisualize the polarity percent
From the calculated polarity we can visualize the polarity
through charts. Compare the polarity with various
personalities and conclude it with the personalities who
tweets maximum positive tweets.
Fig 3.7 Polarity level of Narendra Modi tweet.
Comparison of the tweets polarity of various famous
personalities
From the calculated polarity levels of various
personalities, we can compare their polarity level to find the
polarity of their tweet,whether their tweet is positive or
negative or neutral. Visualize various personalities with their
tweets polarity and compare.
Fig 3.8 Comparison of the Polarity Levels
IV. CONCLUSION
Twitter sentiment analysis comes under the category of
text and opinion mining. It focuses on analyzing the
sentiments of the tweets and feeding the data to a machine
learning model to train it and then check its accuracy, so that
we can use this model for future use according to the results.
It comprises of steps like data collection, text pre-
processing, sentiment detection, sentiment classification,
training and testing the model. This research topic has
evolved during the last decade with models reaching the
efficiency of almost 85%-90%. But it still lacks the
dimension of diversity in the data. Along with this it has a
lot of application issues with the slang used and the short
forms of words. Many analyzers don’t perform well when
the number of classes are increased. Also, it’s still not tested
that how accurate the model will be for topics other than the
one in consideration. Hence sentiment analysis has a very
bright scope of development in future.
REFERENCES
1. Jansen,B.J.; Zhang,M.; Sobel,K.; and Chowdury,A. (2009),
“Twitterpower: Tweets as electronic word of mouth”, Journal of the
American Society for Information Science and Technology
60(11):2169–2188.
2. Pak, A., and Paroubek, P (2010), “Twitter as a corpus for sentiment
analysis and opinion mining”. In Proc. of LREC.
3. Pang, B., and Lee, L. (2008), ”Opinion mining and sentiment
analysis. Foundations and Trends in Information Retrieval” 2(1 -2):1–
135.
4. Wilson, T. Wiebe, J.; and Hoffmann, (P. 2009),”Recognizing
contextual polarity: An exploration of features for phrase-level
sentiment analysis. Computational Li nguistics”, 35(3):399–433.
5. M Hu and B Liu. (2004),”Mining and summarizing customer reviews.
KDD”.
6. L. Barbosa, J. Feng. “Robust Sentiment Detection on Twitterfrom
Biased and Noisy Data”. COLING 2010: Poster Volume,pp. 36-44.
7. J. Kamps, M. Marx, R. J. Mokken, and M. De Rijke, “Using wordnet
to measure semantic orientations of adjectives,” 2004.
8. Alec Go, Richa Bhayani, and Lei Huang. 2009. Twitter sentiment
classification using distant supervision. Technical report, Stanford.
9. David Zimbra, M. Ghiassi and Sean Lee, “Brand-Related Twitter
Sentiment Analysis using Feature Engineering and the Dynamic
Architecture for Artificial Neural Networks”, IEEE 1530-1605, 2016.
10. Varsha Sahayak, Vijaya Shete and Apashabi Pathan, “Sentiment
Analysis on Twitter Data”, (IJIRAE) ISSN: 2349-2163, January 2015.
11. Peiman Barnaghi, John G. Breslin and Parsa Ghaffari, “Opinion
Mining and Sentiment Polarity on Twitter and Correlation between
Events and Sentiment”, 2016 IEEE Second International Conference
on Big Data Computing Service and Applications.
12. Mondher Bouazizi and Tomoaki Ohtsuki, “Sentiment Analysis: from
Binary to Multi-Class Classification”, IEEE ICC 2016 SAC Social
Networking, ISBN 978-1-4799-6664-6.
13. Nehal Mamgain, Ekta Mehta, Ankush Mittal and Gaurav Bhatt,
“Sentiment Analysis of Top Colleges in India Using Twitter Data”,
(IEEE) ISBN -978-1-5090-0082-1, 2016.
14. https://www.geeksforgeeks.org/twitter-sentiment-analysis-using-
python.” Twitter Sentimental analysis for Realdonaldtrump” Devaki
P, Ilakiya, J, Indumathi, R and Arul Priya, M,
15. “Geospatially and literally analysing tweets”, Journal of Advanced
Research in Dynamical and Control Systems Volume 9, Issue Special
Issue 14, 2017, Pages 1002-1009
