Londinium - A MATSim dataset

Londinium is a semi-synthetic dataset for use with MATSim agent based models. The intention is to provide a dataset small enough to use when running MATSim on a typical developer laptop, but big enough to give MATSim a useful workout. Londinium uses real OSM and GTFS data, combined with a synthetic population of agents.

The Londinium dataset is comprised of:

• "pre-baked", ready-made MATSim input files:
  • network
  • vehicles
  • schedule
  • population
• LSOA data - used to generate the agent activity plans
• an OSM data file covering a region in west London - used to generate the MATSim network
• a GTFS public transit schedule for the geographical area covered by the OSM file - used to generate the MATSim network, schedule and vehicles files used as input to MATSim
• a synthetic MATSim population of 100 agents with travel plans inside this region

OSM Data

We chose the location so as to have a selection of different public transit modes, allowing us to see agents interchange modes. We also wanted to have a mixture of major and minor roads, which this particular section of west London gives us.

OSM Files

File	Size	Nodes	Ways	Relations
data/network/osm/Londinium.osm.pbf	4.1M	340,621	65,067	2,129

Geographical Bounds

Population Data

The Londinium synthetic population includes various activities, personal attributes and modes. The agents created were classified into different subpopulations by gender and income. Activities are Home-to-Home round trips, all within 24 hours. Daily activity plans for each agent were created using random choice. Agents choose different travel modes such as bus, walk, car randomly for each trip.

The generation of the Londinium population used a notebook from PAM, our open source population activity modeller

We used the above notebook to generate a population including 100 agents. Each agent has an activity plan for the same 24 hour period and randomly chooses a mode between activities.

Population Files

File	Size
data/population/population.xml	123k

Population Details

No. Agents	100
No. Households	100
No. Subpopulations	3
Subpopulation split	low income: 30% medium income: 40% high income: 30%
Gender split	Female: 50% Male: 50%
Household LAD split	Wandsworth: 26% Kensington and Chelsea: 21% Hammersmith and Fulham: 20% Lambeth: 18% Westminster: 13% Richmond upon Thames: 2%
No. trip legs	352
Leg mode counts	62 mode="bus" 54 mode="car" 50 mode="ferry" 49 mode="subway" 48 mode="walk" 47 mode="rail" 42 mode="bike"
No. Activities	452
Activity type counts	200 "home" 44 "pub" 41 "medical" 40 "shop" 36 "gym" 27 "work" 25 "education" 22 "park" 17 "leisure"

Sample Activity Plans

Population Trips Map

Public Transit Schedule Data

The Londinium PT Network

GTFS Directory

data/network/gtfs/Londinium_arup_gb_rail_and_tnds_2020-03-12

GTFS Files

File	Size
stop_times.txt	93M
trips.txt	7.4M
transfers.txt	56B
stops.txt	69K
shapes.txt	73B
routes.txt	1.1M
frequencies.txt	53B
feed_info.txt	145B
fare_rules.txt	54B
fare_attributes.txt	81B
calendar_dates.txt	6.1M
calendar.txt	714K
agency.txt	77K

GTFS Details

GTFS Day	12 Mar 2020
Transit Trips	98,402
Transit stops	846
Stop times	1,170,122
Transit Routes	20,607
Transit agencies	1,039
Calendar entries	19,541
Calendar dates	386,517

Contributing

Please see our code of conduct and guide to contributing.