Skip to content

ferdamalastofa.md

Latest commit

 

History

History
357 lines (283 loc) · 13.5 KB

ferdamalastofa.md

File metadata and controls

357 lines (283 loc) · 13.5 KB
Raw

Ferðamálastofa (Icelandic Tourist Board)

Inbound tourism statistics via Power BI dashboard scraping. Passenger counts through Keflavík airport by nationality, month, and year.

Data Source

Dashboard URL: https://www.maelabordferdathjonustunnar.is/ Embed backend: ferdapbi.azurewebsites.net (Azure web app generating Power BI embed tokens)

The dashboard embeds Power BI reports via ferdapbi.azurewebsites.net/embed/{reportId}. Data must be extracted by intercepting Power BI API calls within the iframe.

Power BI Report IDs

Report ID Page Description
1fa56a04-3340-46c5-a36b-f9dde4ce0b92 /fjoldi-farthega-um-keflavik Passenger counts by nationality
34af65b4-b68d-4309-a17a-5e9d4632b55c /hotel Hotel guest nights and occupancy
7cf5f866-459c-4873-947c-082d8a216ea9 /allir-gististadir All accommodation types
ae74ab2c-4b1a-4a5c-bf4f-a8ffdda1801f /dvalarlengd-og-gistimati Length of stay and accommodation type

Available Dashboards

Samgöngur (Transport)

Page URL path Description
Þjóðernaskipting og fjöldi um Keflavík /fjoldi-farthega-um-keflavik Primary — passenger counts by nationality
Framboð á flugi frá Keflavík /frambod-a-flugi-fra-keflavik Flight supply from Keflavík
Verð á flugi /verd-a-flugi Flight prices
Tölfræðivísar norrænna flugfélaga /tolfraedivisar-norraenna-flugfelaga Nordic airline statistics
Skemmtiferðaskip /skemmtiferdaskip Cruise ships
Norróna /norrona Norröna ferry

Ferðamenn (Tourists)

Page URL path Description
Aldur, kyn og fleiri bakgrunnsbreytur /aldur-kyn-og-fleiri-bakgrunnsbreytur Age, gender, demographics
Tilgangur ferðar og heimsóttir landshlutar /tilgangur-ferdar-og-heimsottir-landshlutar Purpose of travel, regions visited
Ánægja og upplifun /anaegja-og-upplifun Satisfaction and experience
Fjöldi gesta á áfangastöðum /fjoldi-gesta-a-afangastodum Visitors at destinations
Umferðarslys erlendra ferðamanna /umferdarslys-erlendra-ferdamanna Tourist traffic accidents
Spá um fjölda erlendra farþega /spa-um-fjolda-erlendra-farthega Forecast of foreign passengers

Gisting (Accommodation)

Page URL path Description
Hótel /hotel Hotel statistics
Verð á hótelgistingu /verd-a-hotelgistingu Hotel prices
Allir gististaðir /allir-gististadir All accommodation types
Dvalarlengd og gistimáti /dvalarlengd-og-gistimati Length of stay and accommodation type

Rekstur & efnahagur (Operations & Economy)

(Available but not yet explored)

Key Data: Passenger Numbers (fjoldi-farthega-um-keflavik)

Metrics

Metric Icelandic Description
Heildarfjöldi Total passengers All passengers through Keflavík in period
Fjöldi erlendra farþega Foreign passengers Non-Icelandic passengers (count + %)
Fjöldi Íslendinga Icelandic passengers Icelandic passengers (count + %)
Uppsafnaður heildarfjöldi Cumulative YTD Year-to-date total

Filters/Slicers

Filter Icelandic Values
Markaðssvæði Market area All, Europe, North America, Asia, etc.
Ár Year 2016–2026+
Mánuður Month janúar–desember

Report Tabs

Tab Description
Fjöldi í [mánuður] Monthly breakdown by nationality with YoY comparison
Yfirlit Overview/summary
Árstíðardreifing Seasonal distribution

Nationality Breakdown

Data includes passenger counts by nationality with flags. Top nationalities (Jan 2026 sample): Ísland, Bandaríkin, Bretland, Kína, Ítalía, Þýskaland, Frakkland, Pólland, Suður-Kórea, Ástralía/Nýja-Sjáland, Japan, Spánn, Suður-Ameríka, Kanada, etc.

Key Data: Stays (dvalarlengd-og-gistimati)

Report ID: ae74ab2c-4b1a-4a5c-bf4f-a8ffdda1801f

Tourist length of stay and accommodation type from Ferðamálastofa's border survey (landamærarannsókn).

Metrics

Metric Icelandic Description
Meðalfjöldi gistinátta Average guest nights Monthly average nights stayed, 2024–2026
Sundurliðun eftir bakgrunni Breakdown by demographics Average nights by age group (15–24, 25–34, ..., 65+)
Dvalarlengd Length of stay distribution % of tourists by stay duration (0, 1, 2–3, 4–5, 6–8, 9–12, 13+ days)
Sundurliðun eftir tegund gistingar By accommodation type Usage % and median stay by type

Accommodation Types (tegund gistingar)

Icelandic English Category
Vinir/ættingjar, möbilhúsi… Friends/family, motorhome Ekki greitt fyrir náttuvöl
Tjaldstæði, ekki greitt f… Campsite (unpaid) Ekki greitt fyrir náttuvöl
Önnur gisting Other accommodation Ekki greitt fyrir náttuvöl
Hótel, gistiheimili Hotel, guesthouse Greitt fyrir náttuvöl
Ibúðagisting Apartment rental (Airbnb etc.) Greitt fyrir náttuvöl
Hostel Hostel Greitt fyrir náttuvöl
Tjald greitt fyri… Campsite (paid) Greitt fyrir náttuvöl
Húsbíll/tjaldv… Campervan/motorhome Greitt fyrir náttuvöl
Sumarhús eða skálar Summer houses or huts Greitt fyrir náttuvöl
Húsbíll greitt fyri… Campervan (paid) Greitt fyrir náttuvöl

Filters/Slicers

Filter Icelandic Values
Bakgrunnsbreytur Background variable Aldur (age), Kyn (gender), etc.
Ár Year Multiple selections (2024, 2025, 2026)
Mánuður Month All, or specific months

Sample Output

age_group,avg_nights
15-24 ára,6.8
25-34 ára,7.0
35-44 ára,6.7
45-54 ára,6.8
55-64 ára,6.9
65 ára og eldri,7.2

stay_duration,pct
Gisti ekki,0.6%
1 dagur,2.7%
2-3 dagar,16.8%
4-5 dagar,25.2%
6-8 dagar,30.6%
9-12 dagar,16.0%
13 dagar eða meira,8.2%

Extraction Method

Use Playwright to load the parent page and intercept Power BI executeQueries API calls:

import asyncio
import json
from playwright.async_api import async_playwright

BASE_URL = "https://www.maelabordferdathjonustunnar.is"

async def scrape_passenger_data(page_path="/fjoldi-farthega-um-keflavik"):
    async with async_playwright() as p:
        browser = await p.chromium.launch(headless=True)
        page = await browser.new_page()

        query_results = []

        async def handle_response(response):
            url = response.url
            # Power BI data queries go through these endpoints
            if 'querydata' in url.lower() or 'public/reports' in url.lower():
                try:
                    if response.status == 200:
                        content_type = response.headers.get('content-type', '')
                        if 'json' in content_type:
                            body = await response.json()
                            query_results.append({
                                'url': url,
                                'data': body
                            })
                except Exception:
                    pass

        page.on('response', handle_response)

        await page.goto(
            f"{BASE_URL}{page_path}",
            wait_until='networkidle',
            timeout=60000
        )
        # Power BI reports load async — wait for data queries
        await asyncio.sleep(10)

        await browser.close()
        return query_results

Parsing Power BI Response

Power BI embedded reports use a compressed DSR (DataShapeResult) format:

def parse_powerbi_results(results):
    """Extract tabular data from Power BI query results."""
    rows = []
    for result in results:
        data = result.get('data', {})
        if 'results' not in data:
            continue
        for res in data['results']:
            dsr = res.get('result', {}).get('data', {}).get('dsr', {})
            for ds in dsr.get('DS', []):
                for ph in ds.get('PH', []):
                    dm = ph.get('DM0', [])
                    # Value dictionaries for compressed references
                    value_dicts = dsr.get('ValueDicts', {})
                    for row in dm:
                        # G0, G1, ... = dimension values (nationality, month, etc.)
                        # C = compressed reference indices into ValueDicts
                        # X[n].M0 = measure values
                        # R = repeat flags (inherit from previous row)
                        rows.append(row)
    return rows


def decompress_dsr(dsr_data):
    """Decompress Power BI DSR format with ValueDicts and repeat flags.

    Power BI compresses data by:
    1. ValueDicts: shared string arrays referenced by index (C field)
    2. R (repeat): bitmask indicating which G values repeat from previous row
    3. Ø (null): marks null/missing values
    """
    value_dicts = dsr_data.get('ValueDicts', {})
    all_rows = []

    for ds in dsr_data.get('DS', []):
        for ph in ds.get('PH', []):
            dm = ph.get('DM0', [])
            prev_values = {}

            for row in dm:
                current = {}
                repeat_mask = row.get('R', 0)

                # Resolve G (group/dimension) values
                for i in range(10):  # G0..G9
                    key = f'G{i}'
                    if key in row:
                        current[key] = row[key]
                        prev_values[key] = row[key]
                    elif repeat_mask & (1 << i) and key in prev_values:
                        current[key] = prev_values[key]

                # Resolve C (compressed dict reference) values
                if 'C' in row:
                    for idx, val in enumerate(row['C']):
                        dict_key = f'D{idx}'
                        if dict_key in value_dicts and isinstance(val, int):
                            current[f'C{idx}'] = value_dicts[dict_key][val]
                        else:
                            current[f'C{idx}'] = val

                # Extract X (measure) values
                x_data = row.get('X', [])
                for xi, x in enumerate(x_data):
                    if isinstance(x, dict):
                        for mk, mv in x.items():
                            current[f'X{xi}_{mk}'] = mv

                all_rows.append(current)

    return all_rows

Switching Report Tabs via Interaction

The dashboard has multiple tabs (Fjöldi, Yfirlit, Árstíðardreifing). To load data from different tabs, click the tab buttons:

async def click_tab(page, tab_name):
    """Click a Power BI report tab within the iframe."""
    # Find the Power BI iframe
    iframe_element = await page.query_selector('iframe[src*="ferdapbi"]')
    if not iframe_element:
        return
    frame = await iframe_element.content_frame()
    if not frame:
        return

    # Click the tab button by text
    button = await frame.query_selector(f'button:has-text("{tab_name}")')
    if button:
        await button.click()
        await asyncio.sleep(5)  # Wait for new data to load

Changing Filters (Year/Month)

To scrape different time periods, interact with the Power BI slicers:

async def set_year_filter(frame, year):
    """Change the Ár (year) slicer in the Power BI report."""
    # Find and click the year dropdown
    year_slicer = await frame.query_selector('[aria-label*="Ár"]')
    if year_slicer:
        await year_slicer.click()
        await asyncio.sleep(1)
        option = await frame.query_selector(f'[title="{year}"]')
        if option:
            await option.click()
            await asyncio.sleep(3)

Sample Output

month,year,nationality,passengers,yoy_change,yoy_pct
2026-01,2026,Ísland,43809,-3611,-8%
2026-01,2026,Bandaríkin,24069,-9066,-27%
2026-01,2026,Bretland,20973,+2244,+12%
2026-01,2026,Kína,8908,-764,-8%
2026-01,2026,Ítalía,7555,+5262,+229%
2026-01,2026,Þýskaland,6626,-1260,-16%
2026-01,2026,Frakkland,5803,-115,-2%
2026-01,2026,Pólland,3860,+654,+20%
2026-01,2026,Suður-Kórea,2921,+1304,+81%
2026-01,2026,Ástralía/Nýja-Sjáland,2751,+127,+5%

Data Caveats

  1. Embed token auth: The parent page (maelabordferdathjonustunnar.is) fetches a Power BI embed token from ferdapbi.azurewebsites.net. Direct Power BI API calls require this token.
  2. Rate limiting: Don't scrape too frequently — the embed tokens have limited validity.
  3. DSR compression: Power BI uses a compressed format with ValueDicts, R (repeat) flags, and C (compressed references). Must decompress before use.
  4. Icelandic number format: Uses dots as thousands separators (163.175 = 163,175).
  5. Monthly updates: Data is updated monthly after Isavia publishes new figures.
  6. Historical data: Available from ~2016 onward via year slicer.

Alternative Sources

  • Hagstofan: Has some tourism statistics but less granular nationality breakdown
    • SAM01601: Passengers through Keflavík by nationality (annual)
  • Isavia: Source of the raw data, publishes monthly press releases
  • UNWTO: International comparison data

Save Output

mkdir -p data/processed/ferdamalastofa
# After extraction:
# data/processed/ferdamalastofa/passengers_by_nationality.csv
# data/processed/ferdamalastofa/monthly_totals.csv
# data/processed/ferdamalastofa/stays_by_duration.csv
# data/processed/ferdamalastofa/stays_by_accommodation.csv
# data/processed/ferdamalastofa/hotel_occupancy.csv