Geo-tagged Photos Research Articles

Spatial Analysis of Tourist Spots in Central Tokyo Using Online Geotagged Photographs from Flickr

Abstract. Advancements in information and communication technologies enable us to trace human activities through digital media; networks of fixed or mobile sensors, such as smartphones, global positioning systems, and travel cards, can monitor the timing and locations of human behavior throughout the day. These new sources of geospatial data provide the possibility of analyzing high-resolution big data to capture patterns of human spatial behavior across space and time. Online photo-sharing services are comprised of such sources. This study analyzes the geocoded photos from a photo-sharing service to make a comparison of the varying spatial patterns of tourist attractions in Tokyo as differentiated between those taken by Japanese and foreign visitors. We selected three places (Shinjuku, Asakusa, and Ginza) as study districts from the five top places frequently visited by foreigners based on the 2014 Survey Report of Foreigner Behavior. The data were downloaded from Flickr, one of the most popular online platforms for people to share their travel experiences by uploading photos. We used the data of geocoded photos from 2014, taken within 1 km from Shinjuku, Ginza, and Asakusa stations. Information on the nationality of contributors recorded in the Flickr data was used to compare the tourist spots of Japanese and foreign visitors. To identify the tourist spot, we employed a hotspot analysis using geographic information systems. The obtained results showed some tourist spots common to Japanese and foreign visitors in three study areas, namely, places close to railroad stations. However, the distribution of foreigners’ photographs showed a higher spatial concentration than those taken by Japanese tourists. This tendency is consistent with the findings obtained by Suzuki and Wakabayashi (2008), wherein they compared photos of tourist attractions in guide books of Tokyo. This comparative difference can be explained by the accessibility of railroad stations; in other words, foreigners rely more on public transportation during their comparatively short stays. Spatial patterns of tourist spots varied with the study district in question. In Shinjuku and Ginza, there are differences in tourist spots between Asians and Europeans. Particularly, European tourists’ hot spots tend to be more scattered than Asians. However, such differences are not observed in Asakusa, where tourists’ hot spots are concentrated around Sensoji Temple and Tokyo Sky Tree. The reason for this is that tourist attractions are concentrated in and around specific facilities in the Asakusa district. Hence, the variations of tourist spots obtained by analyzing the geotagged photographs on Flickr can be explained by the difference in behavior and interest of foreign tourists between countries, as well as the spatial distribution of tourist attractions.

Analysis of Tourists’ Image of Seoul with Geotagged Photos using Convolutional Neural Networks

Abstract. In this study we aim to analyze the urban image of Seoul that tourists feel through the photos uploaded on Flickr, which is one of Social Network Service (SNS) platforms that people can share Geo-tagged photos. We first categorize the photos uploaded on the site by tourists and then performed the image mining by utilizing Convolutional Neural Network (CNN), which is one of the artificial neural networks with deep learning capability. In this study we are able to find out that tourists are interested in old palaces, historical monuments, stores, food, etc. in which are considered to be the signatured sightseeing elements in Seoul. Those key elements are differentiated from the major sightseeing attractions within Seoul. The purpose of this study is two folds: First, we analyze the image of Seoul by applying the technology of image mining with the photos uploaded on Flickr by tourists. Second, we draw some significant sightseeing factors by region of attraction where tourists prefer to visit within Seoul.

Analysing tourism movements along the Danube river based on geotagged Flickr photography

Abstract. The Danube, connecting ten countries with large cities, national parks and other smaller destinations is a branded as integrated destination system of Europe. This paper examines the behaviour of visitors along the river by analysing the spatio-temporal distribution of geotagged photos shared on Flickr. The authors downloaded the metadata of more than 2 million geotagged photos along the river, and created a weighed movement graph based on the short time-interval user movements. The network analysis of the graph revealed the central integrating effect of the large capital cities on the entire system. Modularity analysis indicated 18 regional tourism clusters, which are connected into three major separated destination systems which have little relation between them and the isolated clusters of Lower Danube.This paper discusses the method of data aggregation as well as the tools used and created for visualising and analysing the data.

Minding the Gap Between Perceived and Projected Destination Image by Using Information and Communication Platforms and Software

The article examines a double case study framework for analyzing perceived and projected destinations by using ITC solutions. First, 22,362 photos tagged with “Brasov” were collected and analyzed using the Flickr API. Second, a descriptive-explanatory research was employed, applying an instrument for the analysis and to address the online identity of place brands where a proposed online platform generates an automatic score calculation. The spatial patterns of tourist activity revealed many similarities and differences compared to promoted attractions by the DMOs, as the results indicated that geotagged photos reflect the projected image of the destination as the data provided a hotspot distribution of popular tourist attractions. The article makes a theoretical and practical contribution: (a) visual imagery may be more fully implemented in research studies; and (b) the distribution of popular tourist attractions may be in synergy between the perceived and projected image of a destination. Implications for marketing managers are presented.

Identification of Coral Reef Conditions Using Hydro-Acoustic Technology in Giligenting Island, Sumenep Regency Indonesia

Coral reefs are one of the major ecosystems in the coastal region with high biodiversity and resource level. As productive ecosystems with high economic and ecological potential, the existence of coral reef ecosystems must be sustainably managed. One of the areas with high diversity of coral reefs is Giligenting Island located in Sumenep regency. This research was conducted by taking primary data in the form of tidal data, bathymetry data and Side Scan Sonar recordings. From the primary data and then processed until produces 3 dimensional model of seabed features that are then interpreted and analyzed to obtain information about healthy or damaged coral reefs, using underwater geo-tagged photos to validate. The 3-dimensional model has a maximum depth of 7.667 meters and 1.403 meters at minimum depth with an area of 100,536.097 m2. From the interpretation of Side Scan Sonar acoustic images it is found the area of sand feature is 30.79%; the area of mud feature equal to 29.94%; the healthy coral covering 35.75% of the survey area, consisting 10.09% of coral reef barrier reef feature, and 25.66% of small coral reef feature; and 3.52% of damaged coral reef feature from total 43,527.110 m2 Side Scan Sonar survey area.

THE USE OF CROWDSOURCED GEOGRAPHIC INFORMATION FOR SPATIAL EVALUATION OF CULTURAL ECOSYSTEM SERVICES IN THE AGRICULTURAL LANDSCAPE: THE CASE OF CHIANTI CLASSICO (ITALY)

The use of geo-tagged photographs seems to be a promising alternative to assess Cultural Ecosystem Services CESs in respect to the traditional investigation when focusing on the study of the aesthetic appreciation of a protected area or natural landscape. The aim of this study is integrating the cumulative viewshed calculated from geotagged photo metadata publicly shared on Flickr with raster data on infrastructure, historical sites, and the natural environment, using landscape ecology metrics and RandomForest modelling. Crowdsourced data provided empirical assessments of the covariates associated with visitor distribution, highlighting how changes in infrastructure, crops and environmental factors can affect visitor’s use. These data can help researchers, managers, and public planners to develop projects, and guidelines in the rural landscape for incresing the supply for CESs.

Using geotagged photographs and GPS tracks from social networks to analyse visitor behaviour in national parks

ABSTRACTThis article explores the potential of geotagged data from social networks to analyse visitors’ behaviour in national parks, taking the Teide National Park as a study area. Given its unique landscape and characteristics, plus the fact that it is the most visited national park in Spain, Teide National Park presents itself as a suitable candidate to explore new sources of data for studying visitors’ behaviour in national parks. Through data from a social photo-sharing website (Flickr) and GPS tracks from a web platform (Wikiloc), we outline several visitors’ characteristics such as the spatial distribution of visitors, the points of interest with the most visits, itinerary network, temporal distribution and visitors’ country of origin. Additionally, we propose a practical use of geotagged data for determining optimal locations for new facilities such as information stands. Results show that data from social networks is suitable to analyse visitor behaviour in protected areas.

Automatic landmark extraction from geo-tagged social media photos using deep neural network

Wayfinding is one of the most daily activities of peoples in urban spaces which could be facilitated by a set of landmarks to provide a mental map of the environment. Landmark detection in urban spaces is yet one of the great challenges in landmark-based wayfinding. This paper aims to apply new sources of geo-tagged photos derived from social media like Telegram for landmark extraction. This paper proposed a method to extract landmarks automatically through clustering of geo-tagged photos by Density Based Spatial Clustering of Applications with Noise (DBSCAN) method and object detection by deep artificial neural network (deep belief network) algorithm. The proposed method is implemented in 48 different routes of 3 districts of Tehran, Iran. The accuracy assessment demonstrated the efficiency of the algorithms for landmark extraction from huge geo-tagged photographs. Furthermore, the extracted landmarks are evaluated by 120 wayfinders. The experimental results demonstrated the usability of the proposed landmarks in wayfinding process with 87% of user satisfaction.

Strategic Sustainable Development Through the Lenses of Tourists’ Preferences: A Geotagged Photo Approach

Knowing tourists’ preferences and experiences with respect to their national park visits is of great importance to implementing strategically sustainable development of national parks. Flickr geotagged photos are utilized and analyzed as our main data source. We included 7090 photos from 626 people in the Great Smoky Mountains National Park to investigate visitors’ behaviors through the presentations of photo spatial and temporal patterns. The results indicated that tourist behaviors that reflect on what they like and enjoy during their park visits can be extracted from geotagged social media data in terms of frequency and length of enjoyment as visitors’ preferred spots.

Geotagged Photos Clustering Using Adapted Density-Based Spatial Clustering of Applications with Noise Algorithm

Geotagged Photos Clustering Using Adapted Density-Based Spatial Clustering of Applications with Noise Algorithm

Discovering place-informative scenes and objects using social media photos.

Understanding the visual discrepancy and heterogeneity of different places is of great interest to architectural design, urban design and tourism planning. However, previous studies have been limited by the lack of adequate data and efficient methods to quantify the visual aspects of a place. This work proposes a data-driven framework to explore the place-informative scenes and objects by employing deep convolutional neural network to learn and measure the visual knowledge of place appearance automatically from a massive dataset of photos and imagery. Based on the proposed framework, we compare the visual similarity and visual distinctiveness of 18 cities worldwide using millions of geo-tagged photos obtained from social media. As a result, we identify the visual cues of each city that distinguish that city from others: other than landmarks, a large number of historical architecture, religious sites, unique urban scenes, along with some unusual natural landscapes have been identified as the most place-informative elements. In terms of the city-informative objects, taking vehicles as an example, we find that the taxis, police cars and ambulances are the most place-informative objects. The results of this work are inspiring for various fields—providing insights on what large-scale geo-tagged data can achieve in understanding place formalization and urban design.

Big data and evaluation of cultural ecosystem services: an analysis based on geotagged photographs from social media in Tuscan forest (Italy)

Abstract: The paper presents a methodology to quantify the suitability of forest stands for the potential delivery of cultural ecosystem services (CES). The quantification of CES represents a complicated task in the framework of ecosystem service valuation. Compared to traditional investigations, focusing on the study of the aesthetic appreciation of a particular territory, the use of geotagged photographs seems to be a promising alternative to appraise CES. Thus, in order to analyse CES with a particular focus on the aesthetic appreciation of forest stands, this study exploits big data through the analysis of photos shared on the Flickr social network. Crowdsourced datasets are used to depict the geographic location and density of pictures - expressed as the number of photos per unit of surface - as well as their relationship to forest variables and logistic characteristics. The implemented geostatistical model is used to spatialise the results at the regional level (Tuscany forests, Italy). Among the outputs, high values of CES are stressed for high forest and protected areas. From a forest species viewpoint, silver fir, coastal Mediterranean pine, beech and mixed forests seem to be more appreciated compared to other stand typologies such as oaks (e.g., pubescent or Turkey oak) and thermophilic broad-leaved species. Additional quantitative parameters (e.g., elevation, biomass stock and distance to main roads) were significant to the CES assessment. The potential applications of the technique to support forest planning and management are discussed.

Deep Neural Networks and Kernel Density Estimation for Detecting Human Activity Patterns from Geo-Tagged Images: A Case Study of Birdwatching on Flickr

Thanks to recent advances in high-performance computing and deep learning, computer vision algorithms coupled with spatial analysis methods provide a unique opportunity for extracting human activity patterns from geo-tagged social media images. However, there are only a handful of studies that evaluate the utility of computer vision algorithms for studying large-scale human activity patterns. In this article, we introduce an analytical framework that integrates a computer vision algorithm based on convolutional neural networks (CNN) with kernel density estimation to identify objects, and infer human activity patterns from geo-tagged photographs. To demonstrate our framework, we identify bird images to infer birdwatching activity from approximately 20 million publicly shared images on Flickr, across a three-year period from December 2013 to December 2016. In order to assess the accuracy of object detection, we compared results from the computer vision algorithm to concept-based image retrieval, which is based on keyword search on image metadata such as textual description, tags, and titles of images. We then compared patterns in birding activity generated using Flickr bird photographs with patterns identified using eBird data—an online citizen science bird observation application. The results of our eBird comparison highlight the potential differences and biases in casual and serious birdwatching, and similarities and differences among behaviors of social media and citizen science users. Our analysis results provide valuable insights into assessing the credibility and utility of geo-tagged photographs in studying human activity patterns through object detection and spatial analysis.

One net to rule them all: efficient recognition and retrieval of POI from geo-tagged photos

In this work, we present DeepCamera, a novel framework that combines visual recognition and spatial recognition for identifying places-of-interest (POIs) from smartphone photos. Both deep visual features and geographic features of images are explored in our framework. For visual recognition, we first design the HashNet model extended from an ordinary convolutional neural network (ConvNet) by adding a “hash layer” following the last fully connected layer. Furthermore, we compress multiple pre-trained deep HashNets into one single shallow and hash network namely “SHNet” that outputs semantic labels and compact hash codes simultaneously. As a result, it significantly reduces the time and memory consumption during POI recognition. For spatial recognition, a new layer called Spatial Layer is appended to a ConvNet to capture spatial information. Finally, both visual and spatial knowledge contribute to generating a hybrid probability distribution over all possible POI candidates by plugging the spatial layer into SHNet. Notably, the proposed SHNet model can be used for general visual recognition and retrieval. The experiments conducted on real-world datasets and classic datasets (MNIST and CIFAR-10) demonstrate the competitive accuracy and run-time performance of our proposed framework.

Personalized Tour Recommender through Geotagged Photo Mining and LSTM Neural Networks

In this study, a tour recommendation system based on social media photos is proposed. The proposed recommendation system can generate trip tours considering both the user’s current location and interests. First, we exploited the geotagged photo dataset from social media websites, which includes photo related information such as user ID numbers, timestamps, hashtags, and GPS coordinates. With this information, the second step is to group photos and identify those places that could be considered relevant for travellers using clustering algorithms. The third step characterizes the resulting clusters by grouping them into different categories using latent dirichlet allocation (LDA) topic modelling approach. The last step is the generation of tours using a long-short term memory neural network (LSTM). The experiments show that the proposed system can be efficient to advise future travellers about the places they would be more likely to visit and arrange trips for them.

Using social media to identify tourism attractiveness in six Italian cities

The technological revolution and the widespread of social media have allowed people to generate tremendous amounts of data every day. Social networks provide users with access to information. This paper aims to determine the attractiveness of various tourism sites by investigating the behaviour of users through social media. The database involves geo-tagged photos located in six cities serving as a relevant artistic and cultural hub in Italy. Photos downloaded from Flickr, a data-sharing platform. Data analysis was conducted using Mathematica and Machine Learning models approach. The results of our study show maps of the users’ behaviour identify the annual trend of photographic activity in cities and highlight the effectiveness of the proposed methodology that is able to provide with place and user information. The study underline how the analysis of social data can to create a predictive model to formulate tourism scenarios. At the end, general tourism marketing strategies are discussed.

Destination Value Systems: Modeling Visitor Flow Structure and Economic Impact

This study proposes that the structure of visitor flows within a destination significantly influences the overall economic value generated by visitors. In particular, destination network metrics (i.e., density, in-degree centralization, out-degree centralization, betweenness centralization, and global clustering coefficient) for 29 Florida counties were derived from 4.3 million geotagged photos found on the photo sharing service Flickr and then correlated with visitor-related spending reported by the Florida Department of Revenue. The results of regression analyses indicate that density, out-degree centralization, and in-degree centralization are negatively correlated with total visitor-related spending within a destination, while betweenness centralization is found to have a positive relationship. Based on these findings, it is concluded that the economic value generated by tourism is constrained by the destination network structure of supply-side and demand-side interactions. Further, it is argued that a “network orchestrator” approach to management can be used to better manage economic impacts within a destination.

Towards Detecting Social Events by Mining Geographical Patterns with VGI Data

Detecting events using social media data is important for timely emergency response and urban monitoring. Current studies primarily use semantic-based methods, in which “bursts” of certain semantic signals are detected to identify emerging events. Nevertheless, our consideration is that a social event will not only affect semantic signals but also cause irregular human mobility patterns. By introducing depictive features, such irregular patterns can be used for event detection. Consequently, in this paper, we develop a novel, comprehensive workflow for event detection by mining the geographical patterns of VGI. This workflow first uses data geographical topic modeling to detect the hashtag communities with VGI semantic data. Both global and local indicators are then constructed by introducing spatial autocorrelation measurements. We then adopt an outlier test and generate indicator maps to spatiotemporally identify the potential social events. This workflow was implemented using a real-world dataset (104,000 geo-tagged photos) and the evaluation was conducted both qualitatively and quantitatively. A set of experiments showed that the discovered semantic communities were internally consistent and externally differentiable, and the plausibility of the detected events was demonstrated by referring to the available ground truth. This study examined the feasibility of detecting events by investigating the geographical patterns of social media data and can be applied to urban knowledge retrieval.

Quantifying nature-based tourism in protected areas in developing countries by using social big data

Spatial visitation patterns and its features on nature-based tourism are difficult to assess using only a field-based survey, which is costly and labor intensive. However, understanding of a protected area's visitation status is critical, as it can strongly influence the sustainability of natural resources. Hence, it is important to identify ‘where people visit’ and ‘why people visit,’ to evaluate the features attractive to tourists. In this regard, we proposed and applied social big data to investigate nature-based tourism in an ASEAN Heritage Park. Overall, our research was able to effectively illustrate spatial patterns of visitation using 10 years of Flickr geo-tagged photographs. Hotspots of high visitation were identified, while revealing the local spatial impact of distributed attributes. This study offers insights into the applicability of social big data to protected-area management and its potential in reinforcing existing field-based participatory approaches.

Urban Land-Use Classification From Photographs

Land-use (LU) classification of urban areas is conventionally achieved via field survey or remote sensing technologies, which is labor-intensive and time-consuming. With the wide development of social networks such as microblog and ubiquitous network access, images are captured by residents and tourists. In this letter, we propose a method for an automatic urban LU classification using geotagged images from public venues. Our method identifies the LU type depicted in those images that are extrapolated to the local regions bounded by street blocks. Experiments were conducted with geotagged photographs and Open Street Map of an urban area in London, U.K. It was demonstrated that the proposed method achieved overall 76.5% accuracy across five LU types. More importantly, our method demonstrated a greater performance in dealing with a mixture of LU types.