Sustainate

Methodology

How we calculate

Where the data comes from, how trips are compared with flying, and where the numbers are weaker than they look.

This page explains how Sustainate works out journey times, CO₂ figures, best months and rankings. Every CO₂ figure is an estimate for one person travelling one way. The numbers quoted below are the ones the site uses.

Data sources

We build on open data and use it before anything generated.

  • GeoNames: place names, coordinates, countries, regions and populations. It decides which places can be destinations and where they are.
  • Wikidata and Wikipedia: links between our places and their Wikipedia articles, UNESCO World Heritage sites within 10 km of a destination, notable landmarks within 5 km (at most 10 per place) and official websites. Wikipedia page views feed the rankings. We do not republish Wikipedia text.
  • OpenStreetMap: hotels within 5 km of a destination centre (name, star rating and website where mapped). © OpenStreetMap contributors, ODbL. Hotels that are also notable on Wikidata rank first.
  • Wikimedia Commons: photographs. We only use images under CC0, public domain, CC BY or CC BY-SA licences, with an original width of at least 960 px. Every credit is listed on the attributions page.
  • Eurostat: nights spent in tourist accommodation by region and annual air passengers between airport pairs. Both feed the rankings.
  • OurAirports: airport locations and codes.
  • CHELSA: climate data for 1981–2010, giving the average daytime high temperature of every month at each place. It drives the best-month rules.
  • GTFS timetable feeds: published by railways, ferry and coach operators and transport agencies. They give the services between places and their typical journey durations.
  • Natural Earth: coastlines, used when deciding whether a place has a beach.

Where no timetable feed covers a connection, some links were added with the help of a language model and checked against our existing places and routes. Those connections carry estimated rather than timetable-derived times. The introductions and tips on our pages are also written with a language model, working only from a fact sheet of the data above, and are checked before publication. They are not a source of facts.

How routes and times are built

The network is stored as direct hops between neighbouring destinations, for example one train service from one city to the next. Each hop has a mode, an operator where known, a typical journey duration and a CO₂ estimate. A longer journey is built by chaining hops.

To choose between possible chains we use a “sustainable” profile. The search looks for short travel time, adds a small CO₂ term (every 20 kg of CO₂ counts as one hour) and treats each change of train, boat or operator as costing an extra 1 hour, so simpler journeys win. Night trains are allowed.

A journey total is the sum of the durations of its legs plus 20 minutes for each change. Real waits at a change vary and getting to the station is not included, so door-to-door times are longer. Durations are typical values from timetables, not a live schedule.

A route page is only shown when the best sustainable option takes at most 16 hours, or at most 36 hours when the journey includes a ferry. The two places must be at least 150 km apart in a straight line and each must have an airport, so that a flight comparison is possible. A night-train alternative is shown separately when it takes at most 20 hours, only next to a best option of at least 6 hours, at most 6 hours longer than that option, and only when at least 6 hours and 50% of it are spent on night trains.

The flight model

To compare a trip with flying, we estimate a flight between the same two places.

  • Distance: the great-circle distance between the two places. For flight time and for the distance we display, we multiply it by 1.08 to allow for real flights not following the straight line.
  • Door-to-door time: flying time (flown distance divided by a cruise speed of 800 km/h, plus 0.5 hours for taxiing, climb and descent), plus 2.5 hours at the airport (check-in, security, boarding and baggage), plus 1.5 hours for travel between the cities and the airports at both ends.
  • CO₂: great-circle distance multiplied by an emission factor per passenger-kilometre, 0.22928 kg for domestic flights and 0.12786 kg for all others. We use the short-haul factor at every distance, because DESNZ’s short-haul band ends at 3,700 km and every trip on this site is within Europe.

The emission factors come from the UK government’s greenhouse gas conversion factors for company reporting, which we use for the “average passenger” (an average over cabin classes). They already include DESNZ’s 8% allowance for routing, so we apply them to the great-circle distance and do not add our own 1.08 on top. Our 1.08 is used only for time and displayed distance.

Source: UK DESNZ GHG conversion factors 2026 (rev. 31 Jul 2026), Business travel - air, average passenger, with RF: Domestic, to/from UK, Average passenger 0.22928; Short-haul, to/from UK, Average passenger 0.12786 kg CO2e/passenger.km

The factors include radiative forcing. This is DESNZ’s allowance for the additional warming effect of aircraft emissions at altitude (contrails and nitrogen oxides, for example) on top of the CO₂ itself, which is why the figures are quoted as CO₂e. Including it makes flight figures higher than a CO₂-only count, and the size of these effects is scientifically uncertain. Calculators that leave it out will show lower flight emissions and smaller savings.

Flight model constants
Routing uplift (time and displayed distance)×1.08
Cruise speed800 km/h
Taxi, climb and descent0.5 hours
Airport time (check-in to baggage)2.5 hours
Transfers between city and airport, both ends1.5 hours
Domestic flights0.22928 kg CO₂e per passenger-km
All other flights (short-haul factor)0.12786 kg CO₂e per passenger-km

Worked example: a flight of 600 km in a straight line is shown as 648 km. Flying time is about 1.3 hours, door-to-door about 5.3 hours, and the CO₂e is about 76.7 kg per passenger.

CO₂ of trains, ferries and buses

Each hop gets a CO₂ estimate from its mode and its distance, and a journey’s figure is the sum of its hops. The distance is the straight-line distance between the two places of the hop, so the real track or road is somewhat longer.

  • Train: 0.05 kWh of electricity per passenger-kilometre, multiplied by the average grid intensity (kg of CO₂ per kWh) of the countries at the two ends of the hop. Intensities in our table run from 0.03 to 0.65, and 0.2 is used for a country that is not listed. This is a simplified electric-train estimate. It does not know the actual traction (diesel, or a railway’s own power contract) or the countries passed through.
  • Coach: 0.027 kg of CO₂ per passenger-kilometre.
  • Ferry: 0.019 kg of CO₂ per passenger-kilometre for a foot passenger, typical of large ferries. Cars are not included.

The saving against flying is the flight estimate minus the estimate for the best sustainable option, also shown as a percentage of the flight.

Trees: to make a flight’s figure easier to picture, route pages and the planner say how many trees it would take a full year to absorb it. We divide the flight’s CO₂ by 22 kg, the European Commission’s average for what a mature tree absorbs in a year, and round to a whole tree. Young trees absorb much less, so this is an illustration, not an offset.

What “best months” means

We know the average daytime high of every month at each place. A month fits a type of trip when that temperature is in the right band:

  • City break: between 5 and 32 °C.
  • Nature and hiking: between 10 and 28 °C.
  • Sun and beach: at least 17 °C, and the place has a beach nearby.
  • Snow and ski: November to April, a high of 6 °C or lower, and a ski area nearby.

Fitting months are ranked by how close the temperature is to the middle of the band, and the best-ranked months are shown first. This is a temperature rule only. It does not consider rainfall, crowds, opening times, snow depth or events.

How rankings work

Lists are ordered by an estimate of how much travellers look for a place. They are not a judgement of quality.

  • Destinations: a weighted mix of Wikipedia page views over the last twelve months (English and local language), tourist nights in the surrounding region (Eurostat), how many other destinations our network reaches from the place, and a small bonus for rail hubs and UNESCO World Heritage sites. A language-model review then checked the order for travel relevance and could move a place up or down a little, so towns with many page views but little travel interest do not dominate.
  • Routes: pairs with the most air passengers between their airports (Eurostat) come first, then well-known rail and ferry journeys, then pairs among the most popular destinations, preferring distances where flying is a realistic alternative. At most 40 route pages are chosen per destination in one selection run, at most 300 come from the well-known journeys, and the curated set holds about 5,000 pages.
  • Trips that visitors plan in the Sustainate planner can also get a page, for example on the Latest trips list.

How often it is updated

The data is refreshed when the open datasets it comes from are updated and re-imported. We do not follow a fixed schedule. The planner adds places continuously: when someone searches for a place we do not have, it is looked up and added, together with nearby connections. Pages follow the data when it changes.

Limitations

  • Timetables change, and we show typical durations rather than live times. Check the operator before you travel.
  • CO₂ figures are estimates per person for one way. Real emissions depend on the aircraft, how full the vehicle is, how a train is powered and more.
  • Ground CO₂ and distances use the straight line between places, and journey totals count a flat 20 minutes per change and leave out getting to the station.
  • Flight figures use the average passenger and include radiative forcing, as explained above.
  • We do not show seat availability or fares. Any cost band is a rough indication based on journey length, not a price.
  • Coverage is incomplete. A connection missing from the site does not mean it does not exist, and “within 14 hours” statements only count connections in our network.