!pip install wordcloud from wordcloud import WordCloud import matplotlib.pyplot as plt !pip install emoji import pandas as pd import nltk nltk.download('stopwords') from nltk.corpus import stopwords COMMON_WORDS=set(stopwords.words('english')) COMMON_WORDS.update(['mr', 'beast', 'mrbeast', 'jimmy']) COMMON_WORDS.update(stopwords.words('spanish')) COMMON_WORDS.update(stopwords.words('russian')) COMMON_WORDS.update(stopwords.words('french')) import emoji def remove_emoji(string): return emoji.replace_emoji(string, '') remove_emoji('devamı gelsin 🙏devamı gelsin 🙏🥹') def remove_common_words(txt): l = ' '.join(filter(lambda w: not w.lower() in COMMON_WORDS, txt.split())) return l remove_common_words('Xinhan is the most beautiful girlfriend mr aha beast que') !pip install lingua-language-detector

df = pd.read_json('dataset.json') df.head()

cleandf = df.applymap(lambda x: remove_emoji(x) if type(x) == str else x) cleandf

cleandf

supercleandf = cleandf.applymap(lambda x: remove_common_words(x) if type(x) == str else x) supercleandf

from transformers import pipeline # from lingua import Language, LanguageDetectorBuilder # languages = [Language.ENGLISH, Language.FRENCH, Language.GERMAN, Language.SPANISH] # detector = LanguageDetectorBuilder.from_languages(*languages).build() # langs = df['text'].apply(detector.detect_language_of) # langs classifier = pipeline("text-classification", model="j-hartmann/emotion-english-distilroberta-base") classifier(['Hey', 'Hello']) emotions = classifier([x[:100] for x in cleandf['text']]) emotions

df.insert(0, 'emotion', [x['label'] for x in emotions]) df.insert(0, 'emotion_confidence_score', [x['score'] for x in emotions])

df

df.to_csv('output.csv')

import pandas as pd df = pd.read_csv('output.csv') df.head() supercleandf = df.applymap(lambda x: remove_common_words(remove_emoji(x)) if type(x) == str else x) supercleandf

clean_emotion_df = df.drop(df[(df['emotion'] == 'neutral') | (df['emotion_confidence_score'] < 0.7)].index) well = clean_emotion_df[clean_emotion_df['from'] == 'well'] blind = clean_emotion_df[clean_emotion_df['from'] == 'blind'] givemyfriend = clean_emotion_df[clean_emotion_df['from'] == 'givemyfriend'] givesubscriberisland = clean_emotion_df[clean_emotion_df['from'] == 'givesubscriberisland'] last2leave = clean_emotion_df[clean_emotion_df['from'] == 'last2leave']

%matplotlib inline text = ' '.join(supercleandf['text']) # Generate a word cloud image wordcloud = WordCloud(max_font_size=220, background_color="white",width=1920, height=1080).generate(text) wordcloud.to_image().show()

df['from'].unique() clean_emotion_df

for t in clean_emotion_df[clean_emotion_df['emotion'] == 'fear']['text']: if 'help' in t: print(t) # clean_emotion_df[clean_emotion_df['emotion'] == 'fear'] # a = pd.DataFrame() # a['text'] = clean_emotion_df[clean_emotion_df['emotion'] == 'fear']['text'] # a

%matplotlib inline text = ' '.join(clean_emotion_df[clean_emotion_df['emotion'] == 'fear']['text']) # Generate a word cloud image wordcloud = WordCloud(max_font_size=220, background_color="white",width=1920, height=1080).generate(text) wordcloud.to_image().show()

%matplotlib inline text = ' '.join(clean_emotion_df[clean_emotion_df['emotion'] == 'joy']['text']) # Generate a word cloud image wordcloud = WordCloud(max_font_size=220, background_color="white",width=1920, height=1080).generate(text) wordcloud.to_image().show()

%matplotlib inline text = ' '.join(clean_emotion_df[clean_emotion_df['emotion'] == 'surprise']['text']) # Generate a word cloud image wordcloud = WordCloud(max_font_size=220, background_color="white",width=1920, height=1080).generate(text) wordcloud.to_image().show()

%matplotlib inline text = ' '.join(clean_emotion_df[clean_emotion_df['emotion'] == 'anger']['text']) # Generate a word cloud image wordcloud = WordCloud(max_font_size=220, background_color="white",width=1920, height=1080).generate(text) wordcloud.to_image().show()

%matplotlib inline text = ' '.join(clean_emotion_df[clean_emotion_df['emotion'] == 'disgust']['text']) # Generate a word cloud image wordcloud = WordCloud(max_font_size=220, background_color="white",width=1920, height=1080).generate(text) wordcloud.to_image().show()