Demo : https://drive.google.com/file/d/1oxudknzPrwmtIPhfD0SlzmlqOkV8CWKL/view?usp=sharing 🔍 OVERVIEW
This project collects district-level citizen feedback about justice, administration, and local governance, performs sentiment and NLP-based scoring, stores it in a MySQL database, and visualizes it through Flask-powered dashboards using Plotly, Seaborn, and WordCloud.
It integrates Natural Language Processing (TextBlob, Google Translate) for multilingual text analysis, Flask APIs for chatbot interaction, and Power BI-like visualizations for data-driven decision-making.
⚙️ SYSTEM COMPONENTS & WORKFLOW 1️⃣ Frontend (Flask Web App)
The system starts with a web form where users select their district.
Users chat with an interactive justice feedback bot that asks structured questions (e.g., trust, fairness, corruption, accessibility).
Responses are analyzed in real-time.
Files involved:
project.py → Justice System Feedback Chatbot
app.py → District Administration & Justice Feedback Chatbot
templates/index.html → Chat interface (Flask-rendered)
2️⃣ Backend (Flask APIs + NLP Engine)
Both project.py and app.py are Flask servers that handle real-time chat sessions:
🔹 Text Preprocessing & Translation
Uses Google Translator API to detect and translate user messages to English for NLP.
Supports feedback in multiple Indian languages (Tamil, Hindi, etc.).
🔹 NLP & Scoring
Uses TextBlob for sentiment polarity (positive, neutral, negative).
Custom keyword matching improves category-wise scores (trust, responsiveness, corruption, etc.).
Converts qualitative answers into numeric scores (1–5) for analytics.
🔹 Data Storage
Feedback responses are inserted into MySQL tables:
justice_feedback (justice-related)
district_feedback (district-level)
Each entry includes:
District name
Multiple score fields
Overall & sentiment score
Suggestions text
3️⃣ Visualization & Dashboard Layer
Managed by powerbiapp.py.
📊 Dashboard Features:
Pulls feedback from MySQL using pandas.read_sql().
Renders analytical visuals via Plotly, Matplotlib, and Seaborn.
Generates:
Pie chart → Sentiment breakdown
Radar chart → District-wise comparison
Bar chart → Trust & Justice Index
Line chart → Sentiment trends
Word Cloud → Most frequent citizen suggestions
Heatmap → Correlation matrix between justice metrics
Scatter plot → Sentiment vs Trust
“Justice Index” → Weighted average (Trust: 40%, Responsiveness: 30%, Community: 30%)
🌐 Routes:
/ → Main dashboard (all districts)
/ → Filter dashboard for one district
/tamil_heatmap → District heatmap (uses india-districts-727.json for mapping)
4️⃣ Data API Layer
File: load_data.py
Provides lightweight JSON endpoints for front-end or Power BI connectors:
/counts → Returns district-wise feedback counts/average values from MySQL.
Uses CORS so JavaScript or BI tools can fetch data from localhost safely.
5️⃣ Chatbot Training (Optional Module)
File: nlp_analysis.py
Trains a basic ChatterBot using predefined conversations.
Enables local chatbot interaction for AI testing before web integration.
🧩 DATABASE SCHEMA Table: justice_feedback Field Type Description id INT Primary key district VARCHAR District name trust_score FLOAT Citizen trust level responsiveness_score FLOAT Government service response fairness_score FLOAT Legal fairness perception accessibility_score FLOAT Accessibility to justice services corruption_score FLOAT Corruption level indicator community_justice_score FLOAT Local resolution fairness suggestions TEXT Citizen’s improvement suggestions justice_sentiment VARCHAR Sentiment (positive/neutral/negative) overall_score FLOAT Average sentiment score 🧮 EVALUATION METRICS Metric Description Polarity Score Derived from TextBlob sentiment polarity Keyword Score Based on presence of key thematic words Justice Index Weighted index combining key justice indicators Overall Score Average across multiple category scores Sentiment Distribution % Positive, Neutral, Negative feedback 💡 HOW THE DASHBOARD WORKS
Data Ingestion: User feedback stored in MySQL.
Data Fetching: Flask dashboard fetches via SQL queries.
Data Processing: Python libraries (pandas, numpy) clean & transform data.
Visualization: Plotly and Matplotlib generate live interactive charts.
Display: Dashboard HTML templates render visuals dynamically in browser.
Optional Mapping: GeoJSON (india-districts-727.json) creates district heatmaps.
🔗 APIs / External Libraries Used Library Purpose Flask Backend web framework MySQL Connector Database integration TextBlob Sentiment analysis Googletrans Language detection & translation Plotly / Seaborn / Matplotlib Visualization WordCloud Suggestion keyword visualization Pandas / Numpy Data manipulation CORS Secure cross-origin API calls 🎯 OUTCOME
✅ Citizens can share justice-related feedback interactively. ✅ The system automatically quantifies qualitative text. ✅ District administrators can monitor justice perception through dashboards. ✅ The Justice Index helps compare fairness & responsiveness across districts. ✅ Power BI–style analytics (via Plotly) allows decision-makers to improve governance.