Social media images from Peneda-Gerês National Park, Northern Portugal

This dataset contains images from the Peneda-Gerês National Park, Northern Portugal. The images were collected from the Flickr and Wikiloc platforms considering a time period from 2003 to 2017. In respect to the General Data Protection Regulation 2016/679, social media data protected by users’ rights was not downloaded nor analysed. Public data that would potentially contain personal information from social media users was kept anonymous through the study. Data was retrieved through the use of the freely available Flickr’s Application Programming Interface (API), indicating a time window and a bounding box with a pair of coordinates (in our case: minimum latitude: 41.653104; maximum lat.: 42.083595; min. longitude: -8.426270; max. lon.: -7.754076) around Peneda-Gerês. This information was then saved as an excel file with the following attributes: user-id, date taken, latitude, longitude, picture uniform resource locator (url). A first annotation, in the context of cultural ecosystem services (CES), was performed by dividing the photographs of the dataset into “Indoor” and “Outdoor” classes. Only the “Outdoor” pictures were included in this study, since CES are directly connected to nature and environment, which in turn are related to the outside/outdoor. The “Outdoor” images were further divided into two main classes, “Natural” and “Human”, depending on whether the image was dominated by natural or man-made elements. Lastly, a finer annotation for outdoor images was also provided, which encompasses the following six classes: “Species”, “Landscape”, “Nature”, “Human activities”, “Human structures” and “Posing”. “Species” pictures respectively pertained to close-up shots of animals or plants in the wild, translating CES pertaining to biodiversity appreciation. “Landscape” pictures show wide-open shots of nature in the wild, often with a visible horizon most often representing people’s enjoyment of landscape aesthetics. “Human activities” include pictures where people engage in by recreational activities, for instance related to sports such as ski or cycling. “Human structures” include those pictures where man-made structures dominate in the wild, e.g., historical monuments and churches, capturing situations of cultural heritage and spiritual enrichment. “Posing” refers to pictures with people looking at the camera, with recognizable faces, testifying social enjoyment and sense of identity. Finally, “Nature” pictures capture natural elements with no particular feature (such as species) but with an intermediate shot (differing from wide-open shots attributed to landscapes), expressing the appreciation of nature by people.本数据集包含葡萄牙北部佩内达-热尔国家公园（Peneda-Gerês National Park）的影像数据。这些影像采集自Flickr和Wikiloc平台，采集时段为2003年至2017年。 根据《通用数据保护条例2016/679》（General Data Protection Regulation 2016/679），本研究未下载也未分析受用户权利保护的社交媒体数据。研究过程中，对潜在包含社交媒体用户个人信息的公开数据进行了匿名化处理。 数据通过免费开放的Flickr应用程序编程接口（Application Programming Interface, API）检索获得，设置了时间窗口以及围绕佩内达-热尔国家公园的坐标边界框，具体参数为：最小纬度41.653104；最大纬度42.083595；最小经度-8.426270；最大经度-7.754076。 随后将检索到的信息保存为Excel文件，包含以下属性字段：用户ID、拍摄日期、纬度、经度、图片统一资源定位符（Uniform Resource Locator, URL）。 首先基于文化生态系统服务（cultural ecosystem services, CES）的分析框架开展首轮标注，将数据集内的影像划分为“室内”与“室外”两类。由于文化生态系统服务与自然及环境直接相关，而自然与环境均对应户外场景，因此本研究仅纳入“室外”影像。 “室外”影像进一步依据画面主导元素分为“自然”与“人文”两大类别。此外，研究还为室外影像提供了更细致的标注体系，涵盖以下六个类别：“物种”“景观”“自然”“人类活动”“人工设施”与“人像摆拍”。 “物种”类影像指野外动植物的特写镜头，对应与生物多样性欣赏相关的文化生态系统服务。“景观”类影像展现开阔的野外自然场景，通常带有清晰地平线，多体现人们对景观美学的欣赏。“人类活动”类影像包含人们参与休闲活动的画面，例如滑雪、骑行等体育相关场景。“人工设施”类影像以野外的人工建造物为主体，例如历史遗迹与教堂，记录了文化遗产与精神充实的场景。“人像摆拍”类指人物面向镜头且面部可识别的影像，体现了社交愉悦感与身份认同。“自然”类影像则捕捉无特定特征（如物种）的自然元素，采用中景拍摄（区别于景观类的开阔镜头），表达人们对自然的欣赏之情。