dalyo/Estimating-Tourist-Site-Popularity-in-Senegal-Using-Social-Media-Data

# Estimating the Popularity of Tourist Sites in Senegal Using Social Media Data ## Context / Motivation Senegal has a rich touristic heritage (natural, historical, and cultural sites). Analyzing **social media data** provides a way to estimate the *real or perceived popularity* of these sites. This project aims to support: - Tourism authorities in identifying the most attractive and underutilized sites. - Site managers in adapting communication and promotion strategies. - Researchers in understanding digital tourism dynamics. --- ## Project Objectives - Estimate the popularity of Senegalese tourist sites from social media images and metadata. - Apply deep learning methods to classify and recognize tourist sites. - Produce rankings and visualizations of site popularity. --- ## Data Sources - **Images** collected from social media platforms (Instagram, Twitter, Facebook, etc.). - **Metadata**: date, geolocation (if available), interactions (likes, shares, comments). - **Geographic reference data**: official list of Senegalese tourist sites (for validation). --- ## Methodology 1. **Data Collection** – Retrieving images and metadata. 2. **Preprocessing** – Cleaning, filtering, and labeling images. 3. **Learning / Modeling** – - CNN with transfer learning (`VGG16`). - Potential comparison with other architectures (ResNet, EfficientNet, etc.). 4. **Popularity Estimation** – Aggregating results by site (frequency, user interactions). 5. **Visualization** – Graphs, maps, and rankings of the most popular sites. **Key repository files:** - `CNN_transfert_learning_VGG16.py` → CNN training with transfer learning. - `labelisation des images.ipynb` → notebook for labeling images. --- ## Installation / Requirements ### 1. Clone the repository ```bash git clone github.com cd Estimation-de-la-popularite-des-sites-touristiques-au-Senegal-a-partir-de