Skip to content

imadi124/SENTIFY

Repository files navigation

SENTIFY - TURN REVIEWS INTO CLARITY

SENTIFY is a Streamlit app that can either read a CSV file of reviews or fetch Google Play Store reviews by app ID, label each review as Positive, Negative, or Neutral using VADER, and show overall sentiment percentages.

Features

  • Upload a CSV file
  • Select the review text column
  • Fetch Play Store reviews using app ID (no API key)
  • Stratified sampling: Reviews fetched across all star ratings (1-5) for balanced sentiment representation
  • Batch fetch and batch sentiment processing with live progress bars
  • Per-review sentiment label and compound score
  • Overall Positive/Negative/Neutral percentage summary
  • Final verdict based on dominant sentiment
  • Download labeled CSV output with star ratings (Play Store mode)

Tech Stack

  • Python
  • Streamlit
  • Pandas
  • vaderSentiment (VADER)
  • google-play-scraper

Installation

pip install -r requirements.txt

Run

streamlit run app.py

UI Preview

CSV Review Analysis

CSV Analysis Screen

Play Store Review Analysis

Play Store Analysis Screen 1

Play Store Analysis Screen

Input format

  • CSV with at least one column containing review text
  • Choose that column in the app UI
  • For Play Store mode, provide app ID (example: com.whatsapp)

Why Stratified Sampling?

When fetching Play Store reviews, SENTIFY uses stratified sampling by fetching reviews across all 5 star ratings (1-5 stars) instead of just the newest reviews. This ensures:

  • Balanced representation of negative, neutral, and positive reviews
  • More accurate overall sentiment analysis
  • True reflection of the app's actual sentiment distribution

Need for this project

  • Helps app teams monitor user satisfaction continuously from real reviews.
  • Detects post-release problems early (for example, crash complaints or UI friction) before ratings drop.
  • Highlights what users love (such as easy UI or fast performance) to guide product, roadmap, and marketing decisions.
  • Supports growth by improving review quality insights, protecting ratings, and helping teams prioritize high-impact fixes.
  • Aligns with industry-standard review intelligence workflows used by tools like Appbot, Apptweak, and Sensor Tower.

🛠️ Contributing

Contributions to SENTIFY are highly encouraged. Whether it is code improvements, bug fixes, documentation updates, or feature enhancements, feel free to contribute to the project repository.

📜 License

SENTIFY is licensed under the MIT License, granting you the freedom to use, modify, and distribute the code in accordance with the terms of the license.

About

A Streamlit app that analyzes CSV and Google Play Store reviews with VADER to deliver balanced per-review sentiment labels and overall sentiment insights.

Topics

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

 
 
 

Contributors

Languages