home
close
오늘 하루 안보기
Yanolja Attractiveness Index

Yanolja Attractiveness Index: Methodology

The Global Tourism City Attractiveness Index, developed by Yanolja Research, is significant in that it objectively measures the attractiveness of global tourism cities from the perspective of tourists—the actual demand side of tourism.


This index directly reflects how tourists perceive and evaluate the appeal and popularity of tourism cities. It is derived through sentiment analysis and the measurement of positive and negative buzz volumes associated with specific keywords, based on social media data collected in 14 different languages. The analysis covers 191 cities worldwide, offering a comprehensive, data-driven insight into global tourist perceptions.

Category Description
Purpose

Evaluation of tourism city attractiveness
Target cities

191 cities (12 cities max. per country)
Languages
analyzed

14 languages were chosen for social media data analysis based on their large user populations and the maturity of available NLP technologies

- English, Spanish, Arabic, French, Portuguese, Russian, Indonesian, German, Japanese, Turkish, Vietnamese, Korean, Italian, and Thai
Datasources

News and broadcast media were excluded from data collection channels, as they predominantly contain promotional content or issue-driven reporting, which are not aligned with the objectives of this study

- YouTube, Tumblr, Instagram, Blogs, Facebook Public, X(Twitter), Review, Reddit, Facebook, Forums
Timeframe

Annual data aggregation over recent 2 years (2023.6 ~ 2025.5)

- 2024 : 2023.6 ~ 2024.5   - 2025 : 2024.6 ~ 2025.5

- Due to technical limitations, data is only available for the most recent 2 years. Accordingly, each analysis years is operationally defined as the period from June of the previous year to May of the current year
Data collected &
method of analysis

Positive/negative buzz volume by keyword collected and analyzed using AI-based sentiment analysis
Solution used

Brandwatch
① Selection of Tourism Destination Cities

Tourism destination cities were selected based on the following criteria:


  • ① The city appears in the rankings of at least two global city indices.

  • ② The city is located in a country with a population of over 30 million and has a city population of at least 3 million.

  • ③ The city was not included based on the above criteria but was deemed necessary to include based on qualitative judgment.


A maximum of 12 cities per country is selected.

  • Criteria
    Criteria
    Explained
    # of Cities
    Selected
    1. Global City Index

      • Cities ranked in at least 2 of the 12 global city indices

    3. 105 cities
    1. Population

      • Cities in countries with a population of over 30 million and city populations over 3 million

    3. 23 cities(Duplicates with global indices excluded)
    1. Committee Selection(qualitative)

      • Cities not meeting the aforementioned criteria but deemed necessary based on qualitative judgment- At least 1 city per country where possible

    3. 63 cities
Final selection: 191 cities

Selection limited to 12 per country (excluding U.S. and China)

② Language Selection for Analysis

For social media analysis, 14 languages were selected based on large user populations and the availability of advanced NLP technologies, making them suitable for social data analysis.

(Chinese was excluded due to government policies restricting the export of social data from within China)

No. Language Selected
1 English
2 Spanish
3 Arabic
4 French
5 Portuguese
6 Russian
7 Indonesian
8 German
9 Japanese
10 Turkish
11 Vietnamese
12 Korean
13 Italian
14 Thai
③ Data Collection

Brandwatch, a UK-based research company, offers a comprehensive platform for social data collection, processing, analysis, and reporting. It is recognized as one of the leading solutions with the widest data coverage in the world.


In this study, Brandwatch was used to collect social media data. News and broadcast sources were excluded, as most of their content consists of promotional materials or reports on incidents and issues, which were deemed not aligned with the research objectives.

Data
Coverage
Country

  • 229 countries

Language

  • 126 languages

Channel

  • Over 6 million channels

Key
Features

  • Incorporates AI-based LLMs for automated semantic and sentiment classification

  • Widely used by major companies both in Korea and internationally

  • Utilized by the Korea Tourism Organization, indicating strong domain knowledge in the tourism sectorr

brandwatch

A powerful Global Social Suite powered by advanced AI models — from social data analysis to full-scale management, tailored for modern business needs.

  • 1.7T+ social data coverage
  • Official partner of Twitter, Reddit, Meta, Instagram and Tumblr
  • Trusted by over 7,500 global clients
  • Powered by Brandwatch AI Model Iris (with OpenAI GPT)
FACEBOOK Blogs Forums Reviews
X(TWITTER)
INSTAGRAM
YOUTUBE
TUMBLR
④ Construction of Keyword Library

The selection of keywords used for social data collection followed a 4-stage process, resulting in a final set of 419 keywords utilized in the study.

Keyword Library Construction Process
  • [Round 1]
    Keyword Selection

    Initial keyword library built per
    sub-dimension based on expert input

  • [Round 2]
    Keyword Selection
    (Pilot test)

    Pilot test in 5 cities resulted in top 1,000 associative keywords extracted per sub-dimension Keyword Library refined

  • [Round 3]
    Keyword Selection

    Main survey conducted using refined keywords

    Issue-prone keywords excluded during data extraction process
    (e.g) “Weather” excluded due to irrelevant mentions in weather forecasts
    (e.g.) “Bird” triggered noise from unrelated contents like “Angry Birds” or
    “Early Bird” and was filtered using exclusion words

  • [Round 4]
    Keyword Selection
    (Final)

    Keywords with excessively high or low
    frequency removed for analytical validity

Keyword Library: Summary
Dimension Sub-Dimension Keywords
Urban Aesthetics and Natural Scenery Natural scenery and phenomena 30
Flora and fauna 35
Culture and History Historical sites 25
Educational sites 15
Traditional culture 10
Architectural/aesthetic places 8
Religious attractions 35
Experiential Tourism
Contents		 Food 45
Accommodations 10
Shopping 33
Amusement parks 11
Nightlife 27
Sports 20
Activities 30
Festivals/events 22
Hospitality Friendliness of local residents 29
Kindness of service providers 34
Total 419
⑤ Modeling

This index measures tourism city attractiveness based on the positivity ratio derived through sentiment analysis of social data collected using the Keyword Library. Additionally, it measures tourism city popularity by analyzing the buzz volume (i.e., the amount of related mentions).

Tourism City Attractiveness
Attractiveness

Tourism city attractiveness in social data is measured through a qualitative approach using sentiment analysis.


  • Social media data is categorized by dimension based on keywords

  • Sentiment analysis is conducted on each entry to calculate positivity ratio

  • Attractiveness is measured based on the calculated positivity ratio

Reputation

Buzz volume in social data reflects tourists’ voluntary mentions of a given city, indirectly indicating the destination’s visibility and accessibility.


  • Social media data is categorized by dimension based on keywords

  • Buzz volume is calculated for each keyword-based data entry

  • Popularity of tourism cities is measured based on buzz volume

Through this study, Yanolja Research produces two key outcomes: ① the Global Tourism City Attractiveness Index, and ② the tourism city attractiveness rankings.

The Global Tourism Attractiveness Index enables observation of trends in the increase or decrease of attractiveness at both the city and global levels. Meanwhile, the tourism city attractiveness rankings help identify each tourism city’s unique strengths and its relative position within the global tourism landscape.

Global Tourism City
Attractiveness Index

  • Tourism City Attractiveness Index /
    Tourism City Attractiveness Index by Dimension

  • Global Tourism Attractiveness Composite Index /
    Global Tourism Attractiveness Composite Index by Dimension

City-wise Tourism
Attractiveness Ranking

  • Overall Tourism City Rankings /
    Tourism City Rankings by Dimension

  • Regional Tourism City Rankings /
    Regional Tourism City Rankings by Dimension

Attractiveness Tiers by

  • 1st Tier : cities ranked 1st~50th place

  • 2nd Tier : cities ranked 51st ~ 100th place

  • 3rd Tier : cities ranked 101st ~ 150th place

  • Candidate : cities ranked 151th place and below (the rankings among candidate cities will not be disclosed;
    their progress will be monitored over time through continuous observation and analysis)

The Global Tourism City Attractiveness Index by Yanolja Research is measured based on the average attractiveness scores of 191 cities, ranked from 1 to 191. Each index is normalized by converting the 2024 average value to a baseline of 100, and the change in 2025 is calculated relative to this baseline. The measurement of attractiveness for each index is conducted as follows.

Tourism City
Attractiveness
by Sub-Dimension
Positivity Ratio × Buzz Volume Ratio

  • Positivity Ratio = {Positive Buzz÷(Positive Buzz+Negative Buzz)}×100

  • Buzz Volume Ratio = Total mentions for city A within a given sub-dimension
    ÷Total mentions across all sub-dimensions

Tourism City
Attractiveness by
Dimension
Attractiveness Score by Sub-Dimension × Weight per Sub-Dimension

  • The weight of each Sub-Dimension is calculated based on the buzz volume of the corresponding year

Tourism City
Attractiveness
(Overall)
Attractiveness Score by Dimension × Weight per Dimension
Dimension Weight(%)
Urban Aesthetics and Natural Scenery 30.0%
Culture and History 30.0%
Experiential Tourism Contents 30.0%
Hospitality 10.0%

The Global Tourism Attractiveness Composite Index is calculated by summing the overall attractiveness scores of the 191 ranked cities, normalizing the previous year’s total to a baseline of 100, and then measuring the year-over-year change relative to that baseline.