App-Reviews-Sentiment-Analysis

App Reviews Sentiment Analysis is the process of extracting meaningful insights from the textual content of app reviews by determining the sentiment expressed within. Sentiment analysis helps categorize reviews as positive, neutral, or negative, making it easier for developers to prioritize enhancements based on user feedback.

Text PreProcessing

Data statistics

EDA Applied:

Word Cloud: WordCloud for visualizing common words in app reviews.

Box plot

Correlational Heatmap

Histogram

CountPlot:

Finding user Likings based on score ratings

VADER Sentiment Analysis For Positive Sentiments:

VADER Sentiment Analysis for Negative �Sentiment

TextBlob Sentiment Analysis For Positive Sentiments:

TextBlob Sentiment Analysis for Negative Sentiment

VADER is generally more effective because it is specifically tuned for short, sentiment-rich text, which often includes informal language.

Used for:

Social media posts, app reviews, tweets, or other informal text.
Analyze punctuation, capitalization, or emojis as sentiment indicators.

Top Most Positive Review Apps By Vader

Sentiment Analysis Models

Logistic Regression:

Naive Bayes

Support Vector Machine

The sentiment analysis of app reviews was conducted using three machine learning models: Naive Bayes, Logistic Regression, and Support Vector Machine (SVM). Both Naive Bayes and Logistic Regression achieved the highest accuracy of 71.67%, demonstrating their effectiveness in classifying sentiments accurately. Naive Bayes is particularly advantageous due to its simplicity and efficiency for text data, while Logistic Regression offers robust performance with well-separated classes.

Developers can use the sentiment analysis results to:
o Prioritize bug fixes and app improvements based on user sentiment.
o Enhance features that received positive feedback and modify or remove those that were negatively reviewed.
o Monitor app performance continuously by evaluating user feedback in real-time