A linked panel dataset of daily tourism arrivals, sectoral spending, local-unit consumption, and extreme-weather indicators for Jeju Island, South Korea (2019–2025)

Description

This dataset integrates four thematically related but institutionally dispersed data sources for Jeju Island, South Korea, covering January 2019 to September 2025. It was compiled to support a multiscalar analysis of how extreme weather affects tourist arrivals, sectoral spending composition and local spatial consumption patterns. File 1 – Daily tourism arrivals (2,465 observations): Daily visitor counts disaggregated by domestic and international visitors, compiled from the Jeju Tourism Association. Seven dates record zero arrivals, corresponding to weather-induced complete suspension of air and sea access. File 2 – Sectoral tourism spending (486 sector-month observations): Monthly tourism spending in six sectors (retail, transport, food services, accommodation, arts/sports/leisure and other services), in 100 million KRW, sourced from the Korea Tourism Organization Data Lab (datalab.visitkorea.or.kr) and based on Shinhan Card transaction records. In this classification, transport includes land, water and air transport and car rental. File 3 – Local-unit tourism spending (43 administrative units × 81 months): Monthly tourism spending, in 100 million KRW, for 43 sub-municipal administrative units (12 eup/myeon and 31 dong) in Jeju-si and Seogwipo-si, sourced from the Jeju Tourism Organization data platform (data.ijto.or.kr) and based on Shinhan Card transaction records. Subsequent platform restructuring has restricted equivalent access to this historical series. File 4 – Extreme-weather indicators: Daily binary indicators for three extreme-weather types at four Korea Meteorological Administration synoptic (ASOS) stations (Jeju 184, Gosan 185, Seongsan 188, Seogwipo 189): heatwave (daily maximum temperature ≥ 33 °C), heavy rain (daily precipitation ≥ 80 mm) and strong wind (daily maximum wind speed, i.e. the highest 10-minute mean wind speed of the day, ≥ 14 m/s). The dataset spans a pre-pandemic baseline (2019), the COVID-19 disruption period (2020–2022) and the recovery period (2023–2025). Version 2 adds the folder reproduction_code, containing Python code that reproduces all estimates and Fig. 3 of the article “Extreme weather and island tourism at three scales: gateway arrivals, sectoral spending and local redistribution on Jeju Island, South Korea” (submitted to the International Journal of Biometeorology) directly from Files 1–4, together with the resulting output files. See reproduction_code/README.md.

Steps to reproduce

Raw data were collected from four separate institutional sources between 2024 and 2025. File 1: Daily arrival statistics were obtained from the Jeju Tourism Association portal as monthly files (one annual file for 2019; one file per month for 2020–2025, approximately 70 source files) and manually restructured into a long-format daily panel. File 2: Monthly sectoral spending data were downloaded from the Korea Tourism Organization Data Lab (datalab.visitkorea.or.kr) as individual monthly files across 81 months and compiled into a single panel. In this classification, transport includes land, water and air transport and car rental. File 3: Local-unit spending data for 43 administrative units were collected from the Jeju Tourism Organization data platform (data.ijto.or.kr) as individual monthly CSV files downloaded manually across 81 months. Note: subsequent platform restructuring has restricted equivalent access to this historical series. File 4: Daily meteorological records for four synoptic stations (Jeju 184, Gosan 185, Seongsan 188, Seogwipo 189) were downloaded from the Korea Meteorological Administration open data portal (data.kma.go.kr) as annual station files and merged across 2019–2025. Binary threshold indicators were applied: Heatwave (daily maximum temperature ≥ 33 °C), Heavyrain (daily precipitation ≥ 80 mm), Strongwind (daily maximum wind speed, i.e. the highest 10-minute mean wind speed of the day, ≥ 14 m/s, the wind-speed criterion of the KMA strong-wind advisory). Data compilation and initial processing were performed in Microsoft Excel. Statistical analyses for the associated article were first carried out in IBM SPSS Statistics 24. Version 2 adds the folder reproduction_code, which reproduces all estimates and Fig. 3 of the associated article directly from Files 1–4. To run it, keep Files 1–4 in the dataset root, then run “pip install -r requirements.txt” and “python reproduce.py” inside reproduction_code (Python 3.10 or later; tested with Python 3.13, pandas 3.0, SciPy 1.18, statsmodels 0.15 and matplotlib 3.11). The script derives daily weather indicators as the number of stations recording each event (0–4) and monthly indicators as station-days summed over each month; an extreme-weather month is a month with 10 or more combined heatwave, heavy-rain and strong-wind station-days. City-level spending is the sum of the local units in each city. Outputs are written to reproduction_code/results; the files already there were produced by this script. Details are given in reproduction_code/README.md.

Institutions

• Jeju National University

  Jeju-do, Jeju City

Categories

Licence