Sentiment Analysis with Scikit-Learn

sentiment_analysis.py

import pandas as pd
from sklearn.feature_extraction.text import CountVectorizer


text = [
    'Promptly answer to my request.',
    'great turnaround time!',
    'Fast response and great service!',
    'As usual you took the request and completed in as few steps as possible and well done!',
    'Very quick turnaround time and great communication. Thank you!',
    'Really quick turnaround. Thanks!!',
    'reduce the turn around time and increase/upgrade the interaction ability between the client and the pronto team during the building of the website.',
    'Better turn around time for changes and updates. Better copywriting talents for web content rewrite requests.',
    'Get back to us faster',
    'Be more careful when fixing our site to not break the other coding on other pages',
    'All of the information for this listing is already on the website that Pronto built. It is disappointing that this has taken so many tries to get right.',
    'slow communication regarding SEO questions, emailing SEO team direct has been slower that going through my account rep- vague and incomplete answers wasting time',
    'The time difference and language barrier are my main complaints. Otherwise our website looks great.',
    'Process is way too slow. Working with another country seems to really slow things down alot.',
]
label = [
    'positive',
    'positive',
    'positive',
    'positive',
    'positive',
    'positive',
    'neutral',
    'neutral',
    'neutral',
    'neutral',
    'negative',
    'negative',
    'negative',
    'negative',
]
df = pd.DataFrame(data={
    'text': text,
    'label': label
})

vectorizer = CountVectorizer(stop_words='english')

X_train = df.text
y_train = df.label

X_train_vectorized = vectorizer.fit_transform(X_train)

# import
from sklearn import naive_bayes

# instantiate
clf = naive_bayes.MultinomialNB()

# fit
clf.fit(X_train_vectorized, y_train)

# positive, neutral, negative
X_test = [
    'Fast service! Thank you',
    'wish it could have been resolved sooner as it was simple',
    'Cannot get my website where I want because repeated requests are misunderstood, requests are simply not honored.',
]
X_test_vectorized = vectorizer.transform(X_test)

# predict
clf.predict(X_test_vectorized)

# predict
clf.predict(X_test_vectorized)