Evaluating museum visitor experiences based on user-generated travel photos

Huy Quan Vu, Jian Ming Luo, Ben Haobin Ye, Gang Li, Rob LawJournal of Travel & Tourism Marketing 2017 · DOI

Abstract

This research introduces online travel photos published on social media platforms as a complementary data resource to examine the behavior and experience of museum visitors. The practical value of online travel photos is demonstrated through a case study of popular Hong Kong museums, particularly by using the photo content and metadata available from the Flickr platform. The proposed approach is a generic method for understanding museum visitor behavior and preferences, supporting museum practitioners in developing improved products for visitors. The findings are particularly beneficial for tourism managers in promoting and attracting tourists to local museums.

Introduction

Understanding the motivations and behaviors of museum visitors is essential for effective cultural management and tourism promotion. Traditionally, this understanding has relied on surveys, questionnaires, and travel diaries. While valuable, these methods are often resource-intensive, time-consuming, and prone to recall bias. Furthermore, they often fail to capture the visual essence of the visitor experience.

With the advent of social media, visitors now leave "digital breadcrumbs" in the form of geotagged photos and metadata. This study leverages data from Flickr to provide a microscopic view of visitor behavior, analyzing spatial, temporal, and visual dimensions to offer insights that traditional text-heavy methods might miss.

Methodology

The researchers developed a multi-stage analytical pipeline to process and interpret social media data. The process began with Popular Museum Identification, using the Flickr API to scrape photos tagged with "museum" in Hong Kong. To identify the most visited locations, the study employed P-DBSCAN (a density-based spatial clustering algorithm), which accounts for the number of unique visitors to filter out noise from prolific individual posters.

Spatial clustering of museum hotspots in Hong Kong using P-DBSCAN.

Following the identification of hotspots, the data was refined using geographical "bounding boxes" to capture all geotagged photos within a museum's physical footprint, regardless of whether they were explicitly tagged as "museum." The resulting dataset comprised 2,843 photos from 406 unique visitors over a ten-year period (2007–2016).

The analysis was four-fold:

  1. Spatial Analysis: Mapping GPS coordinates to identify visitor clusters.
  2. Temporal Analysis: Examining timestamps to identify peak visiting periods across different demographics.
  3. Content Analysis: Manually labeling photos into categories such as indoor artifacts, outdoor scenes, and selfies.
  4. Multiple Visit Analysis: Tracking unique UserIDs to distinguish between one-time tourists and recurring local visitors.

Results and Discussion

The "Harbor Effect" and Spatial Context

One of the most striking findings was the impact of a museum's physical surroundings on visitor behavior, termed the "Harbor Effect." For institutions like the Museum of Art and the Space Museum, located on the Tsim Sha Tsui waterfront, nearly 50% of the uploaded photos were of the exterior architecture or the Victoria Harbor skyline. In contrast, museums located further inland, such as the History, Science, and Heritage museums, saw over 80% of photos focusing on indoor exhibits and artifacts. This suggests that for waterfront museums, the "place" is as much a draw as the "collection."

Engagement and Photo-Worthiness

The study introduced a "photo-worthiness" metric to measure visitor engagement. While the Museum of Art attracted the highest volume of visitors, the History and Science museums boasted higher engagement per person. Visitors to these institutions took an average of 11.79 photos each, compared to just 1.57 at the Space Museum. This indicates that while some museums are popular "stops," others provide a more immersive experience that encourages active documentation.

Demographic and Temporal Patterns

Visitor patterns varied significantly by origin. International tourists, particularly from the US and UK, showed a strong preference for the Museum of Art, often visiting early in the year (January–April). Local residents and visitors from Mainland China favored the History and Science museums, with peak attendance in July. Furthermore, international visitors were found to be "efficient," often visiting multiple museums within a single month, whereas locals spread their visits over several years, indicating a pattern of recurring leisure rather than singular tourism.

Conclusion

This research demonstrates that social media data can serve as a powerful, unsolicited focus group for cultural institutions. By analyzing visual footprints, museum managers can better align their exhibits and marketing strategies with the actual interests of their diverse visitor base.

Future work should address the inherent sample bias of platforms like Flickr and the "geotag gap" caused by privacy settings. The authors suggest that integrating computer vision and deep learning to automate content analysis will be crucial for scaling this methodology to a global level, providing even deeper insights into the global museum landscape.

Summary updated

Abstract

This research introduces online travel photos published on social media platforms as a complementary data resource to examine the behavior and experience of museum visitors. The practical value of online travel photos is demonstrated through a case study of popular Hong Kong museums, particularly by using the photo content and metadata available from the Flickr platform. The proposed approach is a generic method for understanding museum visitor behavior and preferences, and supports museum practitioners in developing improved products for visitors. The case study findings are particularly beneficial for tourism managers, especially those in Hong Kong, in promoting and attracting tourists to visit local museums.
BibTeX
@article{vu2017evaluating,
  author = {Huy Quan Vu and Jian Ming Luo and Ben Haobin Ye and Gang Li and Rob Law},
  title = {{Evaluating museum visitor experiences based on user-generated travel photos}},
  journal = {Journal of Travel \& Tourism Marketing},
  year = {2017},
  doi = {10.1080/10548408.2017.1363684},
}
Figure