Whatsapp group chat analysis: getting commond words

commondwords.py

commond_words = chat[['Author','Message']].copy()

from nltk.corpus import stopwords
STOPWORDS = stopwords.words('spanish')

stopwords = list(STOPWORDS)
extra = ["<multimedia", "omitido>", "k", "d","si","multimedia", "omitido"]
stopwords = stopwords + extra
commond_words["Message"] = (commond_words["Message"]
                           .str.lower()
                           .str.split()
                           .apply(lambda x: [item for item in x if item not in stopwords])
                           .explode()
                           .reset_index(drop=True)
                 )

commond_words['Message']= commond_words['Message'].apply(remove_emoji)
commond_words['Message']= commond_words['Message'].replace('nan', np.NaN)
commond_words['Message']= commond_words['Message'].replace('', np.NaN)
commond_words['Message']= commond_words.Message.str.replace(r"(a|j)?(ja)+(a|j)?", "jaja")
commond_words['Message']= commond_words.Message.str.replace(r"(a|j)?(jaja)+(a|j)?", "jaja")


words_dict = dict(Counter(commond_words.Message))
words_dict = sorted(words_dict.items(), key=lambda x: x[1], reverse=True)

words_dict = pd.DataFrame(words_dict, columns=['words', 'count'])

fig = px.bar(words_dict.head(10).dropna(), x='words', y='count',
                 labels={'words':'Common Words'}, 
                 height=400)
fig.update_traces(marker_color='#EDCC8B', marker_line_color='#D4A29C',
                  marker_line_width=1.5, opacity=0.6)
fig.update_layout(title_text='Commond Words Chart')
fig.show()