Tourism and World Heritage

From —, — was the fastest-growing region, with international arrivals rising — a year on average. Among countries that started with at least 1 million arrivals, — grew fastest, at — a year.

Growth period

Growth is the average yearly change (compound annual growth rate) in inbound overnight visitors. Both periods end in 2019, the last year before the pandemic.

Fastest-growing region
—

— a year, —

— → — arrivals

Fastest-growing country
—

— a year, —

— → — arrivals

World, for comparison
World

— a year, —

— → — arrivals

Arrivals by region, indexed to — = 100

Average yearly growth by region

Fastest-growing countries

Top 15 by average yearly growth, among countries with at least 1 million arrivals at the start of the period. Smaller places can post huge rates from a tiny base, so they are listed in the table below but left out of this ranking.

World Heritage and tourism

Countries with more World Heritage sites tend to draw more visitors. In 2019 the median country with 21 or more sites received — international arrivals, against — for countries with none. New sites tell a different story: sites inscribed from 2010 to 2019 show — link to tourism growth over the same decade (correlation —).

Sites on the list
—

— cultural · — natural · — mixed · — in danger

Sites vs. visitors
r = —

Correlation between sites inscribed by 2019 and 2019 arrivals, on a log scale, across — countries. 1 would be a perfect match.

Fewest visitors per site
—

— arrivals per site in 2019, across — sites. Among countries with at least 5 sites.

View

Do more heritage sites mean more visitors?

Each dot is a country: sites inscribed by 2019 across, 2019 arrivals up (log scale). The line is the best fit. Hover a dot for details.

Median 2019 arrivals by number of sites

Countries grouped by World Heritage sites inscribed by 2019.

Visitors per site, 2019

Countries with at least 5 sites. High numbers mean heritage sits inside a busy tourism market. Low numbers point to sites that few international visitors reach.

Most per siteSitesArrivals per site
Fewest per siteSitesArrivals per site

Where the sites are

Compare places side by side

International arrivals per year, 1995–2020. Add any country or region; click a chip to remove it.

All countries, —

Sorted by average yearly growth. Uses the region picked above. Greyed rows started below 1 million arrivals.

#CountryRegionStart year arrivals2019 arrivalsYearly growth
About this data

World Bank, World Development Indicators: International tourism, number of arrivals (ST.INT.ARVL), last updated 8 October 2026. Arrivals count overnight visitors who travel to a country other than their usual residence. Original figures come from UN Tourism's Yearbook of Tourism Statistics.

Regions are the World Bank's seven geographic groups. Middle East & North Africa has no regional total for 1995, so it is missing from the 1995–2019 view. Countries need a figure in both the start year and 2019 to be ranked. 2020 is shown in the comparison chart but left out of growth rates: the pandemic cut arrivals sharply and only 132 places reported that year.

World Heritage sites come from the UNESCO World Heritage List export (whc001), 1,273 properties inscribed from 1978 to 2026, with map positions from its coordinates. A site shared by several countries counts once for each. Comparisons with 2019 arrivals use only sites inscribed by 2019. South Sudan, North Korea, the Holy See and Afghanistan have sites but no arrivals figure, so they are not in the comparisons. Correlation shows that two things move together. It does not show that sites cause tourism: large, well-connected countries tend to have both.

Data: World Bank, World Development Indicators (ST.INT.ARVL), updated 8 October 2026 · UNESCO World Heritage List. Built with D3.

How this was built using Claude Code

This dashboard was built in one Claude Code session, through six prompts. Below is each prompt exactly as typed, with a short note on what Claude did in response. Local file paths are shortened to file names.

Get the data

  1. International tourism, number of arrivals (World Bank): data.worldbank.org/indicator/ST.INT.ARVL. Use Download → Excel.
  2. World Heritage List (UNESCO): data.unesco.org/explore/assets/whc001/export. Download the Excel file and, if you want the map layer, the Shapefile (zip).

Transcript

Prompt 1

@"API_ST.INT.ARVL_DS2_en_excel_v2_11289.xls" What is in this data?

Claude opened the World Bank file: international tourist arrivals (overnight visitors) from 1995 to 2020 for 203 countries and 47 regional and income groups, with sparse data for 2020.

Prompt 2

Create simple dashboard that lets users compare how different countries and regions compare in terms of tourism activity. Highlight the fastest grouping place.

Claude computed average yearly growth (CAGR) for 2009–2019 and 1995–2019 and built the dashboard: leader cards, regional index and growth charts, the top 15 countries, a side-by-side comparison and a full table. The fastest-growing region (South Asia) and country (Uzbekistan) are highlighted. A country needs at least 1 million starting arrivals to be ranked, so tiny bases don't dominate.

Prompt 3

@"whc001.zip" @"whc001.xlsx" I just added 2 files: World Heritage sites Excel file and World Heritage sits map file. Can you make the dashboard more informative and insightful by allowing users to discover the relationship between world heritage sites and tourism?

Claude matched UNESCO's ISO country codes to World Bank codes and counted sites per country, using only sites inscribed by 2019 for the 2019 comparison. It then added the World Heritage section: correlation cards, a scatter plot with two views (sites vs. visitors, new sites vs. growth), median arrivals by site count, the most- and fewest-visitors-per-site tables, and a map of every site. Main findings: site count and arrivals correlate at r = 0.66 on a log scale, but sites added from 2010 to 2019 barely track growth (r = 0.12).

Prompt 4

export this to my github and name public repo Tourism and World Heritage Site Nov 2026. Turn on Pages.

Claude rebuilt the dashboard as a standalone static site (HTML, D3 and the two data files), tested it locally, created the public repository hqu/Tourism-and-World-Heritage-Site-Nov-2026 and turned on GitHub Pages.

Prompt 5

I want to let the user see the entire transcript of the prompts I used to build this artifact. Can you add a link "How this was build using Claude Code". It pop ups a page with the transcript. Include the links https://data.worldbank.org/indicator/ST.INT.ARVL and https://data.unesco.org/explore/assets/whc001/export/?flg=en-us in the transcript instructions. so users can download the data file.

Claude added this window, with the links above for downloading the source data, to the GitHub Pages site. It then asked whether to add it to the claude.ai version of the dashboard too.

Prompt 6

Yes, add it

Claude added the same window to the claude.ai version of the dashboard.