A Maltese startup has earned international recognition for using AI to analyse Google Maps photos posted by tourists.
A paper about the project was named the best full paper at the International Federation for IT and Travel & Tourism’s (IFITT) flagship conference ENTER.
Creators of Polzify downloaded thousands of images of Maltese beaches posted by tourists, anonymised them by blurring people’s faces and then fed them into artificial intelligence software.
The AI categorised the images and provided researchers with a high-level overview of what tourists were seeing, sharing and experiencing during their time in Malta.
They say the data could be used by Visit Malta to refine its marketing efforts, by policymakers to assess resources allocated to tourism hotspots and by authorities to quickly identify areas being negatively impacted by overcrowding, littering or construction debris.
The project, VISTA (Visual Interpretation Systems for Touristic Analysis), was developed by Polzify with co-funding through Xjenza Malta’s Research Excellence Programme.
Polzify is a Maltese startup that collects, aggregates and analyses user content about businesses from across the internet, making it easier for business owners to keep track of what is being said about their company across multiple websites – from Google Maps to Tripadvisor or Trustpilot.
Originally designed to focus on text-based content, such as written reviews, the company used the VISTA project to expand its efforts to also include images and videos posted publicly online.
The research paper that resulted from that experiment beat hundreds of other submissions at the ENTER conference to be named the best at the event.
Malta received around four million tourists in 2025, with those visitors spending an estimated €4 billion while in the country, according to data compiled by the National Statistics Office.
But such metrics only capture part of the picture, Polzify CEO and co-founder Matthew Sacco said.
“To truly understand how Malta is perceived as a quality destination, managers need insight into something less tangible – the destination’s image,” he said. That image is reflected in the images they post online.
VISTA researchers gathered almost 8,300 photos of various beaches posted by tourists to Google Maps across 2024.
To safeguard user privacy, researchers anonymised all images by blurring faces before processing them through AI.
That AI analysis found that more than half of all the photos were taken at just five beaches – St Peter’s Pool, the Blue Lagoon, Ghajn Tuffieha, Golden Bay, and Wied il-Ghasri. Unsurprisingly, the bulk of photos were taken during summer months.
Researchers noted that many of those summer snaps showed signs of overcrowding and some revealed problems with litter or debris.
They said the findings could help destination managers align their marketing efforts with authentic visitor experiences, by for instance showcasing lesser-known rocky beaches that are popular with tourists but less advertised.
AI processing of photos and videos uploaded online by tourists could also help authorities monitor littering and overcrowding pressures at key tourist hotspots, they said.
