AI-Powered Travel Planner for Dynamic Itinerary Generation and Optimization

Abstract

Formulating quality itineraries while keeping safety in mind is a challenging task. This difficulty arises from the presence of multiple sources of travel-related information and their constantly changing nature. This challenge becomes even more significant for students and solo travelers, who are often looking for meaningful and personalized travel experiences. Con- sequently, intelligent itinerary generation systems are required to produce personalized and safety-driven travel plans. This paper presents the design of an artificial intelligence- based travel planner that integrates Google Places API, Google Custom Search Engine, web scraping to obtain real-time news and weather data, a machine learning-based classification model, a Spark-based processing layer, and a frontend application. The system consists of two main modules: one responsible for retrieving tourist information, and the other for evaluating safety using structured parameters derived from unstructured dynamic data. The proposed ARRS system is particularly useful for business travelers visiting new locations and for individuals traveling alone, especially in situations where safety awareness is critical. By combining real-time data analysis with machine learning, the system assigns reliable safety scores and generates informed itinerary recommendations.

Introduction

The passage presents a safety-aware AI travel planning system designed to help travelers make real-time, risk-informed itinerary decisions rather than relying on static recommendations.

It identifies a key limitation in existing platforms like Google Maps and TripAdvisor: they provide fixed suggestions without considering dynamic risks such as crime, weather, protests, or user-specific vulnerability (e.g., solo travelers, women, seniors, or budget users). To address this gap, the proposed system introduces a real-time safety scoring and itinerary generation framework.

The system works by:

• Extracting real-time contextual signals (news, weather, alerts) using AI models (e.g., Gemini API)
• Converting unstructured data into structured risk features
• Combining them with location and traveler-profile data
• Applying a Random Forest machine learning model to compute a safety score (safe, moderate, risky)

A key innovation is the weighted safety scoring mechanism, which produces a continuous score rather than just a label, along with rule-based overrides for critical events like extreme weather or protests.

The architecture includes:

• Frontend (Next.js)
• Backend (FastAPI + Node.js)
• ML inference core (Random Forest model)
• External APIs (Google Places, Gemini AI)

The system also supports real-time itinerary generation, prioritizing safer routes and destinations.

Experimental results show that the model achieves reliable classification of travel risk using a dataset of 50,000 samples with 45 features. The preprocessing pipeline includes normalization, encoding, binning, and inference-time safety handling.

Conclusion

An AI-based travel planner for interactive generation and optimization of travel itineraries is presented., we use Gemini AI.To extract real-time safety parameters for the destinationA Random Forest model is trained using this data to classify the destination as Safe, Moderate, and Risky. With the use of 45 heterogeneous features, the trained model reaches an accuracy of 91.4% with a macro F1-score and macro AUC of 0.97 on a dataset of 50,000 samples. The system is implemented as a microservice using FastAPI, while a Next.js serves as the frontend; end-to-end response times are less than 2 seconds. The formulation of weighted safety score extends the classification problem of discrete labels on continuous risk. According to the SHAP analysis, crime rate and human trafficking are the most influential features in this study. The most crucial future task will be real-time and mo- bile deployment. Real-time integration will be enabled from government functionalities like BPRD, MHA, MoRTH and Central Statistical Office (CSO). Advanced techniques will be applied to handle class imbalance.

References

[1] S. Sankar A., N. Kumar K., S. M. Dinesh, S. Abhishek, and A. T. An- jali, “Intelligent trip planning with integrated street view: A seam- less AI-driven approach,” in Proc. 2023 Innovations in Power and Advanced Computing Technologies (i-PACT), pp. 1–6, doi: 10.1109/I- PACT58649.2023.10434354. [2] M. Talakoti, A. Balaram, D. Jagli, N. Deepika, R. L. Priya, and [3] K. Harshini, “Trip Tailor: AI-powered travel planning with itinerary gen- eration and chatbot assistance,” International Journal of Environmental Sciences, vol. 11, no. 4S, pp. 920–929, 2025. [4] T. Primya, G. Kanagaraj, K. S., K. V. K., and N. S., “GenAI- powered itinerary planner,” in Proc. 2025 3rd Int. Conf. Advancements in Electrical, Electronics, Communication, Computing and Automation (ICAECA), pp. 1–8, doi: 10.1109/ICAECA63854.2025.11012327. [5] Y. Chen and C. Zhang, “Personalized travel planning platform based on machine learning and big data analytics,” Journal of Big Data, vol. 5, no. 1, p. 42, 2018. [6] Z. Wang, F. Ye, J. Wang, and Y. Zhang, “An intelligent personalized recommendation model for travel planning,” Journal of Ambient Intelli- gence and Humanized Computing, vol. 11, no. 9, pp. 4277–4290, 2020. [7] A. Botea and M. Mu¨ller, “AI-based trip planning in the travel domain,” in Intelligent Learning Systems and Advancements in Computer Vision, Springer, 2019, pp. 417–435. [8] M. Helmy et al., “Navigating the world with an intelligent tourist guide using generative AI,” in Proc. Int. Telecommunica- tions Conf. (ITC-Egypt), Cairo, 2024, pp. 1–6, doi: 10.1109/ITC- Egypt61547.2024.10620592. [9] S. B. Dasari, V. Vandana, and A. Bhharathee, “Smart travel planner using hybrid model,” in Proc. Int. Conf. Intelligent Data Communication Technologies and IoT (IDCIoT), Bengaluru, India, 2023, pp. 647–652, doi: 10.1109/IDCIoT56793.2023.10053424. [10] A. Kontogianni, E. Alepis, and C. Patsakis, “Promoting smart tourism personalised services via a combination of deep learning techniques,” Expert Systems with Applications, vol. 187, p. 115964, 2022. [11] S. Du, H. Zhang, H. Xu, J. Yang, and O. Tu, “To make the travel healthier: a new tourism personalized route recommendation algorithm,” Journal of Ambient Intelligence and Humanized Computing, vol. 10, no. 9, pp. 3551–3562, 2019. [12] F. Santos, A. Almeida, C. Martins, R. Gonc¸alves, and J. Martins, “Using POI functionality and accessibility levels for delivering personalized tourism recommendations,” Computers, Environment and Urban Sys- tems, vol. 77, p. 101173, 2019. [13] M. S. Ghanim, K. Shaaban, and A. Siam, “An artificial intelligence approach to estimate peak-hour travel time,” in Proc. 2023 Intermountain Engineering, Technology and Computing (IETC), Provo, UT, 2023, [14] pp. 197–202. [15] S. Kumar and A. Suyampulingam, “Comparative evaluation of path planning algorithms in a simulated disaster environment,” in Proc. 2022 IEEE 2nd Mysore Sub Section Int. Conf. (MysuruCon), 2022, pp. 1–7. [16] L. Sadanandan and S. Nithin, “A smart transportation system facili- tating on-demand bus and route allocation,” in Proc. 2017 Int. Conf. Advances in Computing, Communications and Informatics (ICACCI), 2017, pp. 1000–1003, doi: 10.1109/ICACCI.2017.8125972. [17] R. R. Manthena, S. K. Pavuluri, and S. Annamalai, “Route Chat Con- nect: Empowering collaborative travel planning and social connection,” in Proc. 2024 IEEE, 2024. [18] ?I. Topal and M. K. Uc¸ar, “In tourism, using artificial intelligence forecasting with TripAdvisor data,” in Proc. 2018 Int. Conf. Artificial Intelligence and Data Processing (IDAP), Malatya, Turkey, 2018, pp. 1– 5, doi: 10.1109/IDAP.2018.8620874. [19] Y. Zhai, X. Wei, and J. Song, “Design and implementation of a personalized tourism recommendation system based on data mining and collaborative filtering,” Complexity, vol. 2020, pp. 1–13, 2020. [20] X. Cheng, “A travel route recommendation algorithm based on interest theme and distance matching,” EURASIP Journal on Applied Signal Processing, vol. 2021, no. 1, 2021, doi: 10.1186/s13634-021-00759-x. [21] K. H. Lim, J. Chan, S. Karunasekera, and C. Leckie, “Tour recommen- dation and trip planning using location-based social media: a survey,” Knowledge and Information Systems, vol. 60, no. 3, pp. 1247–1275, 2019. [22] E. Park, J. Park, and M. Hu, “Tourism demand forecasting with online news data mining,” Annals of Tourism Research, vol. 90, p. 103273, 2021. [23] F. Dalipi, Z. Kastrati, and T. O¨ berg, “The impact of artificial in- telligence on tourism sustainability: A systematic mapping review,” in Proc. 2023 Int. Conf. Computational Intelligence and Knowledge Economy (ICCIKE), Dubai, UAE, 2023, pp. 119–125, doi: 10.1109/IC- CIKE58312.2023.10131818. [24] L. Zhang, Y. Zheng, and Z. Li, “Deep collaborative filtering for trip recommendation,” in Proc. 2018 World Wide Web Conference (WWW), [25] pp. 671–680, 2018. [26] J. Wu, H. J. Zeng, Z. Chen, and Y. Zheng, “Efficient trip planning query processing on massive trajectory data,” in Proc. ACM SIGMOD Int. Conf. Management of Data, pp. 1353–1368, 2018. [27] K. Huang, B. Cui, X. Zhou, J. Zhang, and X. Wang, “Personalized recommendation for travel packages based on tourist behavior analysis,” IEEE Transactions on Industrial Informatics, vol. 16, no. 7, pp. 4739– 4749, 2020. [28] G. Quattrone, L. Capra, P. De Meo, and E. Ferrara, “Exploiting place features in collaborative filtering for venue recommendations,” User Modeling and User-Adapted Interaction, vol. 25, no. 5, pp. 437–481, 2015. [29] L. Breiman, “Random forests,” Machine Learning, vol. 45, no. 1, pp. 5–32, 2001. [30] J. H. Friedman, “Greedy function approximation: A gradient boosting machine,” Annals of Statistics, vol. 29, no. 5, pp. 1189–1232, 2001. [31] S. M. Lundberg and S.-I. Lee, “A unified approach to interpreting model predictions,” in Proc. 31st Conf. Neural Information Processing Systems (NeurIPS), pp. 4765–4774, 2017. [32] F. Pedregosa et al., “Scikit-learn: Machine learning in Python,” Journal of Machine Learning Research, vol. 12, pp. 2825–2830, 2011. [33] S. Priya and S. Venkatraman, “Comparative approaches in smart travel recommenders,” in Proc. 2023 ICIOT Conf., 2023. [34] A. A. Ali and K. K. Khan, “Adaptive chatbot for intelligent tour guidance,” in Proc. 2023 Int. Conf. AI Applications in Tourism (AICAT), 2023. [35] G. Gutjahr, L. C. Krishna, and P. Nedungadi, “Optimal tour planning for measles and rubella vaccination in Kochi, South India,” in Proc. 2018 ICACCI, pp. 1366–1370, 2018. [36] M. Amiruzzaman et al., “An AI-based framework for studying visual diversity of urban neighborhoods,” Journal of Computational Social Science, vol. 6, no. 1, pp. 315–337, 2023. [37] N. Neshat and S. Moayedfar, “Sustainable tourism with deep learning approaches,” Journal of Eco-Friendly AI Travel, 2021, doi: 10.1080/19407963.2021.1970578.

Copyright

Copyright © 2026 Mr. R. Phani Kishore, Velagapudi Kesini, Thotakura Karthik Surya, Velugoti Tarunya, Shaik Yashif Basha, Pachigolla Gopichand. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.