Ecosystems generate a large amount of value for human well-being. Natural areas such as De Hoge Veluwe National Park in the Netherlands enable a number of recreational activities. The challenge is capturing this value. New sources of crowdsourced geographic information such as the photo-sharing platform Flickr offer a new, rich source of information on how people are interacting with nature. To sort, interpret and map this data, researchers must draw upon deep learning techniques such as computer vision algorithms and natural language processing.
In order to better manage the contributions of ecosystems to human well-being, a better understanding is required of how and where ecosystems are contributing to people’s cultural enjoyment of the landscape. This includes recreation (hiking, cycling, sailing), aesthetic enjoyment and people’s appreciation of biodiversity. A major challenge has been the availability of data. Researchers have generally been limited to small-scale qualitative studies using surveys and questionnaires (Wood et al., 2013).
However, new source sources of crowdsourced geographic information including the vast amounts of data being created via social media promise to provide us with new insights beyond the scope of these traditional methods. On Flickr, millions of people are posting about their daily interactions with nature, using mobile devices to report their location and user profiles to share details about themselves (Di Minin et al., 2015). Its Application Programming Interface provides access to the metadata of all publicly-posted photos including their title, tags, image url, associated user profile and location, accurate up to street level. Researchers have begun using this data to map city characteristics (Zhou et al., 2014) and analyse the content of images for cultural interactions (Richards and Tunçer, 2017).
De Hoge Veluwe National Park in the Netherlands presents an interesting case study to begin identifying recreational activities using this data. The park’s 5,400 hectares of nature receive 540,000 visitors per year. Visitors come to explore the park on foot, by bike and on horseback (Stichting Het Nationale Park de Hoge Veluwe, 2018). Many of these visitors take photos and upload these to Flickr.
Manually sorting this data is not possible so machine learning techniques need to be employed to sort and interpret this data to identify evidence of recreational activities. Computer vision algorithms can identify objects and scenes in images while natural language processing can assist in processing the textual data associated with the images (Najafabadi et al., 2015). The aim of this research will be to employ these methods to identify recreational activities in the park using Flickr data and map these using the image locations.
- To identify recreational activities in De Hoge Veluwe National Park using Flickr imagery and meta-data.
- To map these activities using the locations of relevant images.
- Di Minin, E., Tenkanen, H., Toivonen, T., 2015. Prospects and challenges for social media data in conservation science. Front. Environ. Sci. 3, 63.
- Najafabadi, M.M., Villanustre, F., Khoshgoftaar, T.M., Seliya, N., Wald, R., Muharemagic, E., 2015. Deep learning applications and challenges in big data analytics. J. Big Data 2, 1–21.
- Richards, D.R., Tunçer, B., 2017. Using image recognition to automate assessment of cultural ecosystem services from social media photographs. Ecosyst. Serv. 31, 318–325.
- Stichting Het Nationale Park de Hoge Veluwe, 2018. Annual Report 2017. Hoenderloo.
- Wood, S.A., Guerry, A.D., Silver, J.M., Lacayo, M., 2013. Using social media to quantify nature- based tourism and recreation. Sci. Rep. 3, 1–7.
- Zhou, B., Liu, L., Oliva, A., Torralba, A., 2014. Recognizing city identity via attribute analysis of geo-tagged images. Eur. Conf. Comput. Vis. 519–534.
- Student must have a level of proficiency in Python or R.
- Previous experience with machine learning is preferred.
Theme(s): Human – space interaction, Empowering & engaging communities