CatchmentFour familiar boxes
Let's start simple: walking, cycling, motorised individual transport and public transport. This is the level at which modal split is usually communicated in Central Europe.
The Modal Shuffle
Organising our transport modes in a meaningful way
The transport landscape is becoming more diverse again. Electrification, sharing and digital mobility services create distinctions that hardly mattered in transport surveys twenty years ago. A bicycle is a bicycle — until it is shared with half the city, or a motor helps us to get faster than 25km/h, or a statistician has to decide which branch of the modal split it belongs to. With that many ors, that sounds like a problem of labels. But it is more than that: the rules used to organise transport modes determine what ultimately appears in a modal split and thus effect politics, planing and descision making.
From a transport mode to a modal split
But: a bicycle is not always just a bicycle
Open only the cycling branch. Conventional bicycles and electrically assisted bicycles already show why newer mobility options complicate older categories. And where is the e-bike?
What exactly is car traffic?
Open the motorised individual transport branch. Driver and passenger roles, motorcycles, cars and carsharing sit on different conceptual dimensions.
Keep car traffic fixed and change the survey
Hold the visual focus on the same broad branch and switch from MiD to SrV. What changes is not the traveller, but the classification system.
Follow one category: taxi
Taxi is revealing because physically it may be a car, while statistically it can be assigned to public transport, individual transport or a residual category. And what about Waymo and co.? Isn't autonomous ridehailing the same as autonomous carsharing?
And what about the E-Scooter?
New modes force surveys to decide which established branch they belong to.
Explore the full survey classifications
Sources: MID 2023 Handbuch zur Datennutzung, SrV 2023 Hubrich Methodenbericht, OÜ 2013-2014 Ergebnisbericht
Other mode taxonomies
Tracking systems and modelling software can begin with comparatively flat mode vocabularies.
Google Timeline
User-selectable travel modes mix vehicles, physical activities and contextual activities.
WalkingCyclingDrivingOn a bus
MotionTag
The visible trip-main-mode vocabulary is much closer to the classic survey vocabulary.
WalkBicycleCarBus
MATSim
The software exposes a mode vocabulary and, for analysis, a default main-mode hierarchy.
walkbikecarpt
Our suggestion
The classifications become easier to compare when their different dimensions are separated. We use the person's role as main attribute because this is the closest differentiation to the main modes: pedestrian, rider, driver, passenger. The person's role is garnished with the other attributes below.
For example: A girl is riding a skateboard. which has a small electric motor. The skateboard is borrowed from her brother. She promised to bring it back. As she rides along to the ice cream parlor, she passes a traffic count which registers a pedestrian.
{
"mode_classification": {
"role": "rider",
"vehicle": "skateboard",
"propulsion": "electric support",
"access": "household",
"operating_form": "individual",
"analytical_assignment": "walking"
},
"observation": {
"source": "qualitative_survey",
"source_category": "pedestrian"
}
}The mode becomes: role - rider, vehicle - skateboard, propulsion - electrically supported, access - household, operating form - individually, assignment - walking
Physical vehicle
What physically moves the person? Nothing? bicycle, car, bus, rail, ship, aircraft?
Propulsion
What produces the movement? Human power, electric assistance, battery electric, combustion, hybrid.
Person's role
What role does the observed person have? Driver, passenger, rider, public-transport passenger.
Access or use
How is the vehicle accessed? Owned, household vehicle, shared, rented.
Service or operating form
How is the trip organised? Individually? As scheduled service, taxi, ride-hailing, or demand-responsive?
Analytical assignment
Into which headline group is the observation counted? Walking, cycling, MIV, public transport, other.
MDES hosted at mobilitydatahub.com
Join others working on mobility demand data
With the Mobility Demand Exchange Specification, mobility demand data can be exchanged on various levels: counts, relations, routes and modal shares. It also defines how modal-share information can be represented: as a standalone aggregated mode share, published without requiring the underlying per-mode records. A MECE mode classification would extend this ambition.
The MultiMoFusion-Project
MultiMoFusion addresses a related problem from another direction: different mobility-data sources observe different parts of the same movement. Fusing them therefore also means reconciling what each source can observe, infer and classify. Our research project MultiMoFusion has been analysing mobility demand data with the aim of gaining a higher quality of mode detection - from floating phone data (FPD) on the one hand and from smartphone tracking data (SMASI) on the other.
MultiMoFusion was kicked off in November 24 and will run for three years. The consortium is being funded as part of the FFG call for proposals "Mobilität (2022) - Städte und Digitalisierung". Partners are TU Wien, TU Graz, Invenium and Entwicklungsgesellschaft Wien 3420.
Methodology and sources: the taxonomy comparison is based on the published documentation and code plans of MiD 2023, SrV 2023 and Österreich unterwegs 2013/14, plus the documented/default vocabularies used for the technical systems discussed above. Categories that we split for comparison without a corresponding source category are explicitly marked as revised in the underlying data.