MessVote – Smart Mess Feedback & Analytics System
"Project Statement: MessVote is a smart web-based mess feedback and analytics system that enables students to rate food items and submit textual feedback. The system applies aspect-based sentiment analysis to automatically detect issues related to food quality, hygiene, and service. It provides administrators with real-time dashboards, weekly trend analysis, and structured insights to support data-driven decision-making." "
Problem Statement
Hostel mess management systems lack structured, data-driven feedback mechanisms. Students provide informal ratings and comments, but these are not systematically analyzed. There is no intelligent system to detect recurring complaints, analyze sentiment trends, or identify specific problem areas such as hygiene or food quality. As a result, decision-making remains manual and reactive rather than proactive and analytical.
Literature Review / Market Research
Most existing mess feedback systems collect only numeric ratings without performing deeper textual analysis. Feedback comments are reviewed manually, making it difficult to detect recurring patterns or sentiment trends. Research in sentiment analysis and opinion mining demonstrates effective classification methods, but their integration into hostel mess management remains limited.
Research Gap / Innovation
MessVote integrates automated aspect-based sentiment analysis into a real-time feedback system. Instead of analyzing only star ratings, the system classifies textual feedback into aspects such as food quality, hygiene, and service. The admin dashboard visualizes complaint distribution, performance trends, and weekly reports, enabling structured insights and data-driven decision-making. ?
System Methodology
Dataset / Input
Input includes rating (1–5), textual feedback, mess name, meal type, and timestamp. Preprocessing includes lowercasing, removal of special characters, stopword handling, cleaning empty feedback, and TF-IDF feature extraction.
Model / Architecture
TF-IDF vectorization is applied to textual feedback, followed by a supervised Naive Bayes classifier for sentiment classification (Positive, Neutral, Negative). Aspect detection categorizes complaints into Food Quality, Hygiene, and Service. Results are integrated into a real-time admin dashboard.
Live Execution
VIEW CODE / DEMOResults & Analysis
The proposed system achieved approximately 91.8% classification accuracy on labeled feedback data. Aspect-based analysis successfully identified recurring complaints related to food quality and hygiene. Compared to traditional rating-only systems, MessVote provides structured insights, trend visualization, and automated complaint detection.
Academic Credits
Project Guide
Dr. Anita Shrotriya
Team Member
Prisha Dureja
23FE10CSE00376