sumo offers a wide range of outputs, but one may find it hard to parse and visualize them. Below, you may find some tools that allow to visualize a simulation run's results for being included in a scientific paper. Additional tools read plain .csv-files and were added to the suite as they offer a similar interface.
All these tools are just wrappers around the wonderful matplotlib library. If you are familiar with Python, you must have a look.
The tools share a set of common options to fine-tune the appearance of the generated figures. These options' names where chosen similar to the matplotlib calls.
The tools are implemented in Python and need matplotlib to be installed. The tools can be found in <SUMO_HOME>/tools/visualization.
Current Tools#
Below, you will find the descriptions of tools that should work with the current outputs sumo/sumo-gui generate. To run them, you'll need:
- to install Python
- to install matplotlib
- to set <SUMO_HOME>
All scripts are executed from the command line and you have to give the command line options as listed in the descriptions below. Please note that #common options may be applied to all the scripts listed in the following sub-sections albeit few options may not work for certain scripts.
plotXMLAttributes.py#
Create multiple 2D-plots of 2 arbitrary attributes from on or more xml files aggregated by an third attribute (i.e. detector-id).
Example uses:
python tools/visualization/plotXMLAttributes.py -x x -y y -s fcd.xml
python tools/visualization/plotXMLAttributes.py -x x -y y -s fcd.xml fcd2.xml
The above example draws the paths of all vehicles through the network based on fcd-output. (It is a special case that can also be accomplished with #plot_trajectoriespy)
When option -s is set, a interactive plot is opened that allows identifying data points vehicles by clicking on the plot (dataID is printed on the console)
Option --filter-ids ID1,ID2,... allows restricting the plot to the given data element ids
By setting the special attribute key @RANK then the index of the elements within the input file is used.
Further examples are shown below...
Inductionloop Speed over Time#
Input is inductionloop-output with 30s aggregation from 2 detectors (<e1Detector id="e1Detector_-109_0_0" lane="-109_0" pos="54.06" period="30.00" file="data.xml"/>
Call: python tools/visualization/plotXMLAttributes.py data.xml -x begin -y speed -s
boarding passengers vs delay for each station#
Input is stop-output
Call: python tools/visualization/plotXMLAttributes.py stopinfos.xml -i busStop -x loadedPersons -y delay -s --scatterplot --legend
Fundamental Diagram from edgeData#
Input is edgeData-output with 1-minute aggregation (<edgeData id="example" file="data.xml" period="60"/>)
Call: python tools/visualization/plotXMLAttributes.py data.xml -i id -x density -y left -s --scatterplot --yfactor 60 --ylabel vehs/hour
Each color gives encodes a different edge-id. Option --factor 60 is used to convert from vehicles per 60s (edgeData-period 60) to vehicles per hour.
plot_trajectories.py#
Create plot of all trajectories in a given --fcd-output file. This tool in particular is located in <SUMO_HOME>/tools.
Example use:
python tools/plot_trajectories.py fcd.xml -t td -o plot.png -s
The option -t (--trajectory-type) supports different attributes that can be plotted against each other. The argument is a two-letter code with each letter encoding an attribute that is derived from the fcd input.
Available Attributes#
- t: Time in s
- d: Distance driven (starts with 0 at the first fcd datapoint for each vehicle). Distance is computed based on speed using Euler-integration. Set option --ballistic for ballistic integration.
- a: Acceleration
- s: Speed (m/s)
- i: Vehicle angle (navigational degrees)
- x: X-Position in m
- y: Y-Position in m
- k: Kilometrage (requires --fcd-output.distance)
- g: gap to leader (requires --fcd-output.max-leader-distance)
Examples Trajectory Types#
- td: time vs distance
- ts: time vs speed
- ta: time vs acceleration
- ds: distance vs speed
- da: distance vs acceleration
- xy: Spatial plot of driving path
- kt: kilometrage vs time (combine with option --invert-yaxis to get a classic railway diagram).
Interactive Plot#
When option -s is set, a interactive plot is opened that allows identifying vehicles by clicking on the respective line (vehicle ids is printed in the console).
Filtering#
Option --filter-route EDGE1,EDGE2,... allows restricting the plot to all trajectories that contain the given set of edges.
Option --filter-ids ID1,ID2,... allows restricting the plot to the given vehicle ids
plot_net_dump.py#
plot_net_dump.py shows a network, where the network edges' colors and width are set in dependence to defined edge attributes. The edge weights are read from "edgedumps" - edgelane traffic, edgelane emissions, or edgelane noise.
 |
It shows the shift in traffic in the city of Brunswick, Tuesday-Thursday week type after establishing an environmental zone. |
 |
Showing the according changes in NOx emissions. |
- both, --dump-inputs <FILE>,<FILE> as well as --measures <STRING>,<STRING> expect two entries being divided by a ','. The first is used for the edges' color, the second for their widths. But you may simply skip one entry. Then, the default values are used.
- dump-files cover usually more than one interval. To generate an extra output file for each interval, use the string '%s' as part of the output filename (this part will be replaced with the corresponding begin time).
Caution
If two input files are given they must cover the same time intervals or no data will be plotted.
Options
Here the most important options are listed. Use --help to see all options.
| Option | Description |
|---|---|
| -n <FILE> --net <FILE> |
Defines the network to read |
| -i <FILE>,<FILE> --dump-inputs <FILE>,<FILE> |
Defines the dump-output files to use as input |
| -m <STRING>,<STRING> --measures <STRING>,<STRING> |
Define which measure to plot;default: speed,entered |
| -w <FLOAT> --default-width <FLOAT> |
Defines the default edge width; default: .1 |
| -c <COLOR> --default-color <COLOR> |
Defines the default edge color |
| --min-width <FLOAT> | Defines the minimum edge width; default: .5 |
| --max-width <FLOAT> | Defines the maximumedge width; default: 3 |
| --log-colors | If set, colors are log-scaled |
| --log-widths | If set, widths are log-scaled |
| --min-color-value <FLOAT> | If set, defines the minimum edge color value |
| --max-color-value <FLOAT> | If set, defines the maximum edge color value |
| --min-width-value <FLOAT> | If set, defines the minimum edge width value |
| --max-width-value <FLOAT> | If set, defines the maximum edge width value |
| -v --verbose |
If set, the progress is printed on the screen |
| --internal | If set, internal edges (of junctions) are included to the genrated shapes. |
plot_net_selection.py#
plot_net_selection.py reads a road network and a selection file as written by sumo-gui. It plots the road network, choosing a different color and width for the edges which are within the selection (all edge with at least one lane in the selection).
 |
The example shows the selection of an "environmental zone". |
Options
| Option | Description |
|---|---|
| -n <FILE> --net <FILE> |
Defines the network to read |
| -i <FILE> --selection <FILE> |
Defines selection file to read |
| --selected-width <FLOAT> | Defines the width for selected edges;default: 1 |
| --selected-color <COLOR> | Defines the color for selected edges; default: r (red) |
| --edge-width <FLOAT> | Defines the default edge width; default: .2 |
| --edge-color <COLOR> | Defines the default edge color; default: #606060 |
| -v --verbose |
If set, the progress is printed on the screen |
plot_net_speeds.py#
plot_net_speeds.py reads a road network and plots it using the allowed speeds to color the edges. It is rather an example for using measures read from the network file.
 |
The example colors the streets in Brunswick, Germany by their maximum allowed speed. |
Options
| Option | Description |
|---|---|
| -n <FILE> --net <FILE> |
Defines the network to read |
| --edge-width <FLOAT> | Defines the default edge width; default: .2 |
| --edge-color <COLOR> | Defines the default edge color; default: #606060 |
| --minV <FLOAT> | Define the minimum value boundary |
| --maxV <FLOAT> | Define the maximum value boundary |
| -v --verbose |
If set, the script says what it's doing |
plot_net_trafficLights.py#
plot_net_trafficLights.py reads a road network and plots it and additionally adds dots/markers at the position of traffic lights that are part of the net.
 |
The example shows the traffic lights in Brunswick. |
Options
| Option | Description |
|---|---|
| -n <FILE> --net <FILE> |
Defines the network to read |
| --edge-width <FLOAT> | Defines the default edge width; default: .2 |
| --edge-color <COLOR> | Defines the default edge color; default: #606060 |
| --minV <FLOAT> | Define the minimum value boundary |
| --maxV <FLOAT> | Define the maximum value boundary |
| -v --verbose |
If set, the script says what it's doing |
plot_summary.py#
plot_summary.py reads one or multiple summary-files and plots a selected measure (attribute of the read summary-files). The measure is visualised as a time line along the simulation time.
 |
The example shows the numbers of vehicles running in a large-scale scenario of the city of Brunswick over the day for the standard week day classes. "mo.xml", "dido.xml", "fr.xml", "sa.xml", and "so.xml" are summary-files resulting from simulations of the weekday-types Monday, Tuesday-Thursday, Friday, Saturday, and Sunday, respectively. |
Options
| Option | Description |
|---|---|
| -i <FILE>[,<FILE>]* --summary-inputs <FILE>[,<FILE>]* |
Defines the summary-file(s) to read |
| -m <STRING> --measure <STRING> |
Defines the measure to read from the summary file; default: running |
| -v --verbose |
If set, the progress is printed on the screen |
| --dpi <FLOAT> | Define dpi resolution for figures; default: None |
plot_tripinfo_distributions.py#
plot_tripinfo_distributions.py reads one or multiple tripinfo-files and plots a selected measure (attribute of the read tripinfo-files). The measure is visualised as vertical bars that represent the numbers of occurrences of the measure (vehicles) that fall into a bin.
|
The example shows the travel time distribution for the vehicles of different week day classes (Braunschweig scenario). "mo.xml", "dido.xml", "fr.xml", "sa.xml", and "so.xml" are tripinfo-files resulting from simulations of the weekday-types Monday, Tuesday-Thursday, Friday, Saturday, and Sunday, respectively. |
Options
| Option | Description |
|---|---|
| -i <FILE>[,<FILE>]* --tripinfos-inputs <FILE>[,<FILE>]* |
Defines the summary-file(s) to read |
| -m <STRING> --measure <STRING> |
Defines the measure to read from the summary file |
| -v --verbose |
If set, the progress is printed on the screen |
| --bins <INT> | The number of bins to devide the values into |
| --norm <FLOAT> | Defines a number by which read values are divided; default: 1.0 |
| --minV <FLOAT> | The minimum value; if set, read values that are lower than this value are set to this value |
| --maxV <FLOAT> | The maximum value; if set, read values that are higher than this value are set to this value |
plot_csv_timeline.py#
plot_csv_timeline.py reads a .csv file and plots columns selected using the --columns <INT>[,<INT>]* option. The values are visualised as lines.
 |
The example shows the New European Driving Cycle (NEDC). |
Options
| Option | Description |
|---|---|
| -i <FILE> --input <FILE> |
Defines the input file to use |
| -c <INT>[,<INT>]* --columns <INT>[,<INT>]* |
Defines which columns shall be plotted |
| -v --verbose |
If set, the progress is printed on the screen |
plot_csv_pie.py#
plot_csv_pie.py reads a .csv file that is assumed to contain a name in the first and an according value in the second column. The read name/value-pairs are visualised as a pie chart.
|
Note: Please note that you should set the width and the height to the same value using the --size <FLOAT>,<FLOAT> option, see #common options. Otherwise you'll get an oval. |
Options
| Option | Description |
|---|---|
| -i <FILE> --input <FILE> |
Defines the input file to read |
| -p --percentage |
Interprets read measures as percentages |
| -r --revert |
Reverts the order of read values before plotting them |
| --no-labels | Does not plot the labels |
| --shadow | Puts a shadow below the circle |
| --startangle <FLOAT> | Sets the start angle |
| -v --verbose |
If set, the progress is printed on the screen |
plot_csv_bars.py#
plot_csv_bars.py reads a .csv file that is assumed to contain a name in the first and an according value in the second column. The read name/value-pairs are visualised as a bar chart.
|
|
Options
| Option | Description |
|---|---|
| -i <FILE> --input <FILE> |
Defines the csv file to use as input |
| -x <INT> --column <INT> |
Defines which column of the read .csv-file shall be plotted; default: 1 |
| -r --revert |
Reverts the order of read values before plotting them |
| -w <FLOAT> --width <FLOAT> |
Defines the width of the bars; default: .8 |
| --space <FLOAT> | Defines the space between the bars; default: .2 |
| --norm <FLOAT> | Defines a number by which read values are divided; default: 1.0 |
| --show-values | Shows the values |
| --values-offset <FLOAT> | Position offset for values |
| --vertical | Draws vertical bars (default are horizontal bars) |
| --no-labels | Does not plot the labels |
| -v --verbose |
If set, the progress is printed on the screen |
common options#
The following options are common to all previously listed tools. They can be divided into two groups:
- options for formatting the figure (including adding captions etc.)
- options for determining what to do with the generated figure
The options are listed and discussed in the following sub-sections, respectively.
Formatting Options#
| Option | Description |
|---|---|
| --colors <COLOR> | Uses the given colors; the number of given colors must be the same as the number of measures to plot |
| --colormap <STRING> | Uses the named colormap |
--labels <LABELS>-l <LABELS> |
Uses the given labels; the number of given labels must be the same as the number of measures to plot |
--xlim <XMIN>,<XMAX> |
Describes the limits of the figure along the x-axis |
--ylim <YMIN>,<YMAX> |
Describes the limits of the figure along the y-axis |
--xticks <XMIN>,<XMAX>,<XSTEP>,<XSIZE>--yticks <XSIZE> |
If only one number is given, it is interpreted as the size of the tick labels on the x-axis; if four numbers are given, they are interpreted as the lowest ticks position, the highest ticks position, the step between ticks, and the tick's size, respectively, all along the x-axis |
--yticks <XMIN>,<XMAX>,<XSTEP>,<XSIZE> --yticks <XSIZE> |
If only one number is given, it is interpreted as the size of the tick labels on the y-axis; if four numbers are given, they are interpreted as the lowest ticks position, the highest ticks position, the step between ticks, and the tick's size, respectively, all along the y-axis |
| --xtime1 | If set, the tick labels along the x-axis are formatted as time entries (hh:mm) |
| --ytime1 | If set, the tick labels along the y-axis are formatted as time entries (hh:mm) |
| --xtime2 | If set, the tick labels along the x-axis are formatted as time entries (hh:mm:ss) |
| --ytime2 | If set, the tick labels along the y-axis are formatted as time entries (hh:mm:ss) |
| --xgrid | If set, a grid along the ticks on the x-axis is drawn |
| --ygrid | If set, a grid along the ticks on the y-axis is drawn |
| --xticksorientation <FLOAT> | The orientation of the x-ticks (in °); default: matplotlib default |
| --yticksorientation <FLOAT> | The orientation of the y-ticks (in °); default: matplotlib default |
| --xlabel <STRING> | Defines the label to set for the x-axis |
| --ylabel <STRING> | Defines the label to set for the y-axis |
| --xlabelsize <FLOAT> | Defines the size of the label of the x-axis |
| --ylabelsize <FLOAT> | Defines the size of the label of the y-axis |
| --title <STRING> | Defines the title of the figure |
| --titlesize <FLOAT> | Defines the size of the title |
| --adjust <FLOAT>,<FLOAT> --adjust <FLOAT>,<FLOAT>,<FLOAT>,<FLOAT> |
Adjust the plot; If two floats are given, they are interpreted as left and bottom values, if four numbers are given, they are interpreted as left, bottom, right, top |
| --size <FLOAT>,<FLOAT> | Defines the size of figure |
| --no-legend | If set, no legend is drawn |
| --legend-position <STRING> | Defines the position of the legend; default: matplolib default |
Interaction Options#
If one of the scripts is simply started with no options that are listed below, the figure will be shown. To write the figure additionally into a file, the filename to generate must be given using the --output <FILE> (or -o <FILE> for short).
If the script is run in a batch file, it is often not convenient to show the figure (once known it is as it should be). In such cases, the option --blind (-b) can be used that suppresses showing the figure.
| Option | Description |
|---|---|
| -o <FILE> --output <FILE> |
Defines the name under which the figure shall be saved |
| -b --blind |
If set, the figure will not be shown |
Further Visualization Methods#
Coloring edges in sumo-gui according to arbitrary data#
sumo-gui can load weight files and show their values when setting edge coloring mode to by loaded weight. When stepping through the simulation, different time intervals contained in the weight file can be shown.
sumo-gui -n NET --weight-files FILE --weight-attribute ATTR -e 86400
Suitable weight files are those produced by edgeData-output. To show the number of departed vehicles for each edge, the option --weight-attribute departed would be used.
The weight files generated by randomTrips option --weights-output-prefix can also be used (to visualize depart/arrival probabilties).
Intersection Flow Diagram#
To visualize the flow on an intersection with line widths according to the amount of traffic, the tool route2poly.py can be used.
- Use netedit to select all edges at one or more intersection for which the flow shall be visualized and save the selection to a file (e.g. sel.txt)
-
generate polygons with widths according to the number of vehicles passing the selected edges. When setting --scale-width to 0.01, 100 vehicles using the same edge sequence will correspond to a polygon width of 1m. The option --spread is used to prevent overlapping of the generated polygons and should be adapted according the the polygon width.
route2poly.py NET routes.rou.xml -o flows.poly.xml --filter-output.file sel.txt --scale-width 0.01 --internal --spread 1 --hue cycle -
visualize the flows in sumo-gui
sumo-gui -n NET -a flows.poly.xml
Outdated#
The tools are meant to be named as following: <API>__<DESCRIPTION>.py, where:
- <API>: mpl for matplotlib
- : the SUMO-output that is processed (mainly)
- <DESCRIPTION>: what the tool does
mpl_dump_twoAgainst.py#
Reads two dump files (mandatory options --dump1 <FILE> and --dump2 <FILE>, or, for short -1 <FILE> and -2 <FILE>). Extracts the value described by --value (default: speed). Plots the values of dump2 over the according (same interval time and edge) values from dump1.
Either shows the plot (when --show is set) or saves it into a file (when --output <FILE> is set).
You can additionally plot the normed sums of the value using (--join). In the other case, you can try to use --time-coloring to assign different colors to the read intervals.
You can format the axes by using --xticks <XMIN,XMAX,XSTEP,FONTSIZE> and --yticks <YMIN,YMAX,YSTEP,FONTSIZE> and set theit limits using --xlim <XMIN,XMAX> and --ylim <YMIN,YMAX>. The output size of the image may be set using --size <WIDTH,HEIGHT>.
mpl_tripinfos_twoAgainst.py#
Reads two tripinfos files (mandatory options --tripinfos1 <FILE> and --tripinfos2 <FILE>, or, for short -1 <FILE> and -2 <FILE>). Extracts the value described by --value (default: duration). Plots the values of tripinfos2 over the according (same vehicle) values from tripinfos1.
Either shows the plot (when --show is set) or saves it into a file (when --output <FILE> is set).
You can format the axes by using --xticks <XMIN,XMAX,XSTEP,FONTSIZE> and --yticks <YMIN,YMAX,YSTEP,FONTSIZE> and set theit limits using --xlim <XMIN,XMAX> and --ylim <YMIN,YMAX>. The output size of the image may be set using --size <WIDTH,HEIGHT>.
mpl_dump_timeline.py#
Reads a value (given as --value <VALUE>, default speed) for edges defined via --edges <EDGEID>[,<EDGEID>]* from the dumps defined via --dumps <DUMP>[,<DUMP>]*. Plots them as time lines, using the colors defined via --colors <MPL_COLOR>[,<MPLCOLOR>]*. Please note that the number of colors must be equal to number of edges * number of dumps.
Either shows the plot (when --show is set) or saves it into a file (when --output <FILENAME> is set).
You can format the axes by using --xticks <XMIN,XMAX,XSTEP,FONTSIZE> and --yticks <YMIN,YMAX,YSTEP,FONTSIZE> and set theit limits using --xlim <XMIN,XMAX> and --ylim <YMIN,YMAX>. The output size of the image may be set using --size <WIDTH,HEIGHT>.
mpl_dump_onNet.py#
Reads a network (defined using --net-file <NET> or -n <NET>) and an edge-dump file (--dump <DUMPFILE> or -d <DUMP_FILE>). Plots the network using the geometries read from <NET>. Both the width and the colors used for each edge are determined using --value <WIDTHVALUE>,<COLORVALUE> where both <WIDTHVALUE> and <COLORVALUE> are attributes within the dump-file that exist for each edge.
You can change the used color map by setting --color-map <DEFINITION>. <DEFINITION> is made of a sorted list of values (between 0 and 1) and assigned colors. This means that the default 0:#ff0000,.5:#ffff00,1:#00ff00 let streets with low value for <COLORVALUE> appear red, for those in the middle yellow and for those with a high value green. For values between the given values, the color is determined using linear interpolation. Please note that only lowercase hexadecimal characters may be used.
Either shows the plot (when --show is set) or saves it into a file (when --output <FILENAME> is set).
--join sums up the values found for each edge and divides the result by the number of these values. If join is not set and --output is given, one should choose an output name which looks as following: <NAME>'%05d.png. The %05d will be replaced by the current time step written.
If you have generated a set of images by not "joining" (aggregating) the data, you can convert the obtained pictures into an animated gif using ImageMagick and the following command:
convert -delay 20 *.png -loop 0 animation.gif
(loop 0 means that the animation repeats from begin after the end)
You can format the axes by using --xticks <XMIN,XMAX,XSTEP,FONTSIZE> and --yticks <YMIN,YMAX,YSTEP,FONTSIZE> and set theit limits using --xlim <XMIN,XMAX> and --ylim <YMIN,YMAX>. The output size of the image may be set using --size <WIDTH,HEIGHT>.