For getting MATSim outputs into SimWrapper

In general, most MATSim outputs are too big and use unwieldy formats. The scripts below can help convert some MATSim outputs into smaller, more internet-friendly formats.

If you develop your own scripts, please let us know and we can add them here!

Create a JSON network appropriate for loading into SimWrapper visualizations

• Download script here: create-json-network.py

Command:

python create-json-network.py [network] [coord-system]

Inputs: MATSim network.xml.gz file; coordinate system

Outputs: network.json.gz which loads into SimWrapper much faster than an .xml.gz file

• Download script here: create-csv-network.py

Create a CSV network appropriate for loading into R with the sfnetworks package

Command:

python3 create-csv-network.py [my-network.xml.gz]

Inputs: MATSim network.xml.gz file

Outputs: network.csv file with one row per link. Includes a wkt column with the WKT LINESTRING field, from which an R sfnetwork can be created.

Sample code to read the output CSV into R:

library(tidyverse)
library(sfnetworks)
library(sf)

sf <- st_as_sf(read_csv("my-network.csv"), wkt="wkt", crs=25832)
network <- as_sfnetwork(sf)

ggplot() +
  geom_sf(data=st_as_sf(filtered, "edges"), col="grey50") +
  geom_sf(data=st_as_sf(filtered, "nodes"), aes(size=1)

Parse the event file containing DRT events.

• Download script here: parse-drt-link-events.py

Command: python parse-drt-link-events.py [network] [events] [coord-system]

Inputs: network.xml.gz file; events.xml.gz file; a valid coordinate system

The command will run much faster if you filter the events file to only contain drt events first. You can use grep or a similar tool, with a command like:

• zcat events.xml.gz | grep "drt" > drt-events.xml
• But that strips the xml header and footer so actually doesn't work. You need the <events> and </events> tags. I'll fix this soon..?

Outputs: drt-vehicles.json

Use gzip to compress that output so things load faster.