Preserving Heritage – Enhancing Tourism with AI
This capstone project addresses a dual challenge: preserving historical heritage through automation and enhancing the tourist experience through intelligent recommendations. It is divided into two core parts involving both deep learning and machine learning techniques.
Project Objectives
1. Heritage Structure Classification with TensorFlow 🧠
Using TensorFlow, I developed a deep learning model to classify historical structures based on images. Leveraging transfer learning with a pre-trained CNN backbone, the model was fine-tuned to predict the category of heritage structures (e.g., temples, forts, monuments). Key steps included: - Image preprocessing and augmentation - Transfer learning using frozen convolutional layers - Fine-tuning the top dense layers with dropout regularization - Model training with validation accuracy monitoring - Visualization of training metrics to detect overfitting
2. Tourist Recommendation Engine with EDA 📊
For the second part, I used machine learning and EDA to build a personalized tourist recommendation engine. This included: - Data cleaning and exploratory data analysis on user demographics and location preferences - Insights on the most visited and highly rated tourist places - Development of a collaborative filtering model to suggest attractions based on user behavior
Datasets Used
Structures_dataset.zip– Images of historical structuresDataseet_test– Images of historical structures for validationuser.csv,tourism_with_id.csv,tourism_rating.csv– User demographics and tourism activity
Outcome
The combined AI solutions automate heritage categorization and intelligently connect tourists with culturally rich destinations, supporting both preservation and sustainable tourism.