import pandas as pd tweets = pd.read_csv('vaccination_tweets.csv')

tweets.head()

tweets.columns

source_counts = tweets.groupby('source').size().to_frame().reset_index().rename(columns = {0:'count'})

top_sources = source_counts.sort_values(by='count', ascending=False).head(10) top_sources

import seaborn as sns import matplotlib.pyplot as plt sns.barplot(data = top_sources, x = 'source', y = 'count'); plt.xticks(rotation = 'vertical');

country_counts = tweets.groupby('user_location').size().to_frame().reset_index().rename(columns = {0:'count'}) top_countries = country_counts.sort_values(by='count', ascending=False).head(3) sns.barplot(data = top_countries, x = 'user_location', y = 'count'); plt.xticks(rotation = 'vertical');

tweets['text'] = tweets['text'].apply(lambda text: str(text)).dropna()

tweets['text'] = tweets['text'].apply(lambda text: str.lower(text))

import re punctuations = r'[^a-zA-Z0-9]'

tweets['text'] = tweets['text'].apply(lambda text: re.sub(punctuations, " ", text))

tweets['text'].head()

from textblob import TextBlob

TextBlob('covid is terrible i lost my job and i am going crazy at home.').sentiment.polarity

TextBlob('i so happy and excited to get my vaccine').sentiment.polarity

TextBlob('the vaccine is being administered to my pharmacy').sentiment.subjectivity

TextBlob('i think the vaccine is scary the side effects are bad for me.').sentiment.subjectivity

tweets['polarity_score'] = tweets['text'].apply(lambda x: TextBlob(x).sentiment.polarity).values

sns.histplot(data=tweets, x='polarity_score', stat='density', binwidth=0.1); plt.title('Distribution of Polarity Score');

tweets.sort_values(by='polarity_score', ascending=False).head(10)['text']

sns.scatterplot(x=tweets['favorites'], y=tweets['polarity_score']); plt.title('Favorites Vs. Polarity Score');

def score_to_description(polarity): if polarity > -1 and polarity < -0.5: return 'Very Negative' elif polarity > -0.5 and polarity < 0: return 'Negative' elif polarity > 0 and polarity < 0.5: return 'Positive' elif polarity > 0.5: return 'Very Positive' else: return 'Neutral' tweets['polarity_classification'] = tweets['polarity_score'].apply(score_to_description)

sns.histplot(data=tweets, x='polarity_classification', stat='density'); plt.title('Distribution of Polarity Classification'); plt.xticks(rotation = 'vertical');

tweets[tweets['polarity_classification']=='Very Negative']['text'].head(5)

tweets['subjectivity_score'] = tweets['text'].apply(lambda x: TextBlob(x).sentiment.subjectivity).values

sns.histplot(data=tweets, x='subjectivity_score', stat='density', binwidth=0.1); plt.title('Distribution of Subjectivity');

tweets.sort_values(by='subjectivity_score', ascending=False).head(10)['text']

india_tweets = tweets[tweets['user_location']=='India'] malay_tweets = tweets[tweets['user_location']=='Malaysia']

plt.hist(data=india_tweets, x='polarity_score', density=True, alpha=0.4, label='India'); plt.hist(data=malay_tweets, x='polarity_score', density=True, alpha=0.4, label='Malaysia'); plt.legend(); plt.title('Title');

pfizer = tweets[tweets['text'].str.contains('pfizer')] moderna = tweets[tweets['text'].str.contains('moderna')]

plt.hist(data=pfizer, x='polarity_score', density=True, alpha=0.4, label='Pfizer'); plt.hist(data=moderna, x='polarity_score', density=True, alpha=0.4, label='Moderna'); plt.legend(); plt.title('Title');