TufanTicket --- localhost:80085
Revolutionizing event feedback with AI-driven sentiment analysis and actionable insights for seamless event planning.
Created on 22nd March 2025
•
TufanTicket --- localhost:80085
Revolutionizing event feedback with AI-driven sentiment analysis and actionable insights for seamless event planning.
The problem TufanTicket --- localhost:80085 solves
Tagline
Revolutionizing event feedback with AI-driven sentiment analysis and actionable insights for seamless event planning.
The problem it solves
Planning large-scale events often lacks structured, real-time feedback mechanisms, leaving organizers with fragmented data from multiple sources. Our platform bridges this gap by providing AI-powered sentiment analysis and granular feedback on events from genres like music concerts, comedy shows, and public gatherings in India. Users can search for any event and instantly access insights such as positive/negative sentiment percentages, feedback on Content Quality, Venue & Facilities, Staff & Service, and Value for Money.
Event organizers can understand public perception better, address concerns, and improve future events. For instance, data-driven insights like “poor crowd management at X concert” or “stellar sound quality at Y festival” help in immediate decision-making and long-term strategy building.
Our tool also pulls data from multiple sources—user feedback, public reviews, and news—aggregating them into a single dashboard with actionable suggestions for improvement. This empowers organizers to enhance attendee satisfaction, making events safer, more efficient, and enjoyable.
Challenges we ran into
One major challenge was aggregating feedback from diverse data sources like social media, news, and CSV datasets while maintaining data consistency. To overcome this, we built a robust ETL (Extract, Transform, Load) pipeline that cleanses and unifies data into a common format.
Another challenge was achieving accurate sentiment classification with limited labeled data. We tackled this by using pre-trained transformer models fine-tuned on event-specific data to improve precision. Balancing real-time feedback analysis with processing speed was another hurdle; we implemented asynchronous processing and optimized our backend with FastAPI and caching mechanisms.
Deploying our solution in real time via Ngrok, especially on Google Colab, posed restrictions in maintaining persistent connections. We handled this by integrating with Ngrok’s authenticated tunnels and automated reconnections for smooth frontend-backend communication.