Introduction
The Marathi Travel Chat Dataset is a comprehensive collection of over 12,000 text-based conversations between customers and call center agents. Focused on real-life travel and tourism interactions, this dataset captures the language, tone, and service dynamics essential for building robust conversational AI, chatbots, and NLP solutions for the travel industry in Marathi-speaking markets.
Participant & Chat Overview
•
Participants:
200+ native Marathi speakers from the FutureBeeAI Crowd Community
•
Conversation Length:
300–700 words per chat
•
Turns per Chat:
50–150 dialogue turns across both participants
•
Chat Types:
Inbound and outbound
•
Sentiment Coverage:
Includes positive, neutral, and negative interaction outcomes
Topic Diversity
The dataset encompasses a wide range of travel and tourism use cases across both customer-initiated and agent-initiated conversations:
•Booking assistance and travel planning
•Destination information and recommendations
•Flight delays or cancellations
•Lost or delayed baggage support
•Assistance for travelers with disabilities
•Health and safety travel inquiries
•Promotional offers and travel package deals
•Booking confirmations and schedule updates
•Flight change notifications
•Customer satisfaction surveys
•Visa expiration and renewal reminders
•Loyalty and feedback collection campaigns
This variety ensures wide applicability in both sales enablement and customer support automation.
Language Diversity & Realism
Conversations are crafted to reflect the everyday language and nuances of Marathi-speaking travelers:
•
Naming Patterns:
Marathi personal names, airline and hotel names, tour operators
•
Localized Details:
Regional email formats, phone numbers, locations, and cultural references
•
Time and Currency Expressions:
Dates, local times, and prices represented in Marathi forms
•
Slang and Informal Speech:
Common phrases and idioms used in travel planning and customer support
These linguistic and cultural cues enable the development of context-aware, natural-sounding AI systems.
Conversational Structure & Flow
The dataset captures a variety of interaction types, including:
•Quick inquiries and confirmations
•Complex issue resolution
•Advisory and planning sessions
•Travel disruption and recovery support
•Greetings and authentication
•Information request and validation
•Problem or request resolution
•Status updates and follow-ups
•Feedback and escalation (where applicable)
This structure enables training of intelligent dialogue systems that can adapt to dynamic, multi-turn travel conversations.
Data Format & Structure
Available in TXT, CSV, and JSON formats, each record includes:
•Turn-based speaker labeling
•Participant identifiers (anonymized)
•Optional metadata for topic, sentiment, or region
The data is structured for compatibility with standard NLP frameworks and pipelines.
Applications
This dataset can power a wide array of Marathi-language travel AI and NLP use cases:
•Travel Assistant Chatbots
•Multilingual Travel NLP Research
•Booking Automation & Inquiry Handling
•Sentiment & Intent Detection
•Entity Recognition for Dates, Locations, Prices
•Text Summarization and Language Generation
•Customer Experience Personalization Models
Secure and Ethical Collection
•
Informed Consent:
All data contributors participated voluntarily with written consent
•
Privacy Compliant:
No personally identifiable information is shared
•
Secure Storage:
All data was processed and retained within FutureBeeAI’s secure infrastructure
•
Ethical Practices:
Collection and usage align with AI responsibility and data protection standards
Updates & Customization
The dataset is actively maintained and can be tailored to suit your specific project needs:
•
Custom Annotations:
Sentiment tags, NER labels, intent classification, etc.
•
Topic Expansion:
Collect new data for specific segments like visa services, cruise bookings, or hotel support
•
Regional Customization:
Marathi dialect variations based on country or customer segment
•
Multilingual Options:
Equivalent datasets available in other languages upon request
Licensing
This dataset is developed and owned by FutureBeeAI and is available under commercial licensing. Flexible licensing models are available for enterprise, research, and startup use.
