From: Measuring sustainable tourism with online platform data
Publication | Main variable | Data Source | Data type | Ground truth | Country/Region | Sample size |
---|---|---|---|---|---|---|
Bassolas et al. (2016) [39] | Touristic site attractiveness | Geolocated tweets and user home locations | – | Africa, Asia, Europe, North and South America | 9.6 million geolocated tweets; 59,000 users’ places of residence | |
Batista e Silva et al. (2018) [32] | Average daily numbers of overnight tourists | Booking.com; TripAdvisor; Eurostat | Accommodation location and capacity | – | EU-28 (incl. GB) | 716,103 establishments |
Buning and Lulla (2020) [28] | Spatio-temporal usage patterns of bike shares | Bike fleet GPS; user zip codes | GPS; User data | – | Indianapolis, USA | 353,733 individual trips |
Falk and Hagsten (2020) [34] | Visitor flows to world heritage sites | Instagram; UNESCO | Number of posts, hashtags | – | Europe, North America | 680m Instagram posts for 525 sites |
Fatehkia et al. (2020) [42] | Wealth Index at clustered geographic locations | Advertisement market audience size estimates | DHS wealth index | India and Philippines | 1,205 (Philipines); 28,043 (India) | |
Gallego and Font (2021) [30] | Air passenger demand forecasts | Skyscanner flight searches (ForwardKeys) | Global air capacity and flight search data | – | Global | 5,000m searches and >600m picks |
Grybauskas et al. (2021) [44] | Apartment revisions | Property listing sites | Property listings with up to 16 numerical features | – | Vilnius, Lithuania | 18,922 listings |
Hardy and Arval (2020) [29] | Tourist movements in national park | Mobile app; GNSS | Location data; Demographic survey data | – | Tasmania, Australia | 472 tourists (4-14 days, 1 signal/10 seconds) |
Kashyap and Verkroost (2021) [43] | LinkedIn GGI = gender gaps along different dimensions | Advertisement market audience size estimates | Country-level professional gender gaps data from international labour organisation (ILO) | Global, up to 234 countries predicted (depending on level of analyis); Up to 185 in ground truth | 460 million users, 165,02 million without missing data | |
Londoño and Hernandez-Maskivker (2016) [38] | Customer feedback to GreenLeader program | TripAdvisor | Sustainability mention, gender, nationality, hotel category, city, GL level | Review concerning sustainability or not | 6 Cities in Europe and North America | 572 comments |
Mariani and Borghi (2021) [37] | eWOM (presence and depth of discourse) | Booking.com; TripAdvisor | Text (Comments) | – | Americas, Europe | 4,12 million TripAdvisor and 1,56 million Booking.com reviews |
Mendoza et al. (2019) [41] | Intensity of an earthquake in terms of damages | Tweets (text mining) | Earthquake catalogue provided by Seismological Center of Chile | Chile | Initially 825,310 tweets; final sample: 187,317 geo-mapped tweets | |
Nurmi et al. (2020) [31] | Nights spent by foreign tourists | GDS; Amadeus, Sabre, Galileo | Flight bookings | Official accommodation statistics (Finland) | Finland | 58 months × 7 countries |
Quattrone et al. (2018) [33] | Number of Airbnb listings per tract (geographical unit based on census data) | Airbnb | Geolocations of listings | – | 8 cities in USA | 54,681 listings |
Serrano et al. (2021) [36] | Attributes important to “green tourists” | Airbnb | Text (Comments) | Actual user assessments | Global, 83 cities | more than 176 million comments |
Sun and Paule (2017) [35] | Restaurant and bar rating clusters | Yelp | User ratings from 1-5 | – | Pheonix, USA | 2578 restaurants, 981 fast food restaurants, 797 bars |
Talebi et al. (2021) [40] | Potential to become ecotourism destination | Manually collected geo data | Geo data (slope, elevation, soil texture, vegetation, etc.) | Classification from systemic analysis | Arasbaran, Iran | 637 recreational areas |