SSS10 Proceedings of the 10th International Space Syntax Symposium

The space syntax toolkit:

Integrating depthmapX and exploratory spatial analysis
workflows in QGIS
Jorge Gil
Space Syntax Laboratory, The Bartlett School of Architecture, UCL
[email protected]

Tasos Varoudis
Space Syntax Laboratory, The Bartlett School of Architecture, UCL
[email protected]

Kayvan Karimi
Space Syntax Laboratory, The Bartlett School of Architecture, UCL
[email protected]

Alan Penn
Space Syntax Laboratory, The Bartlett School of Architecture, UCL
[email protected]

Abstract
The Space Syntax Toolkit is a QGIS plug-in for spatial network and statistical analysis. It provides a
front-end for the depthmapX software within QGIS, offering user-friendly space syntax analysis
workflows in a GIS environment. It is primarily aimed at supporting the space syntax methodology,
and enhancing it with GIS data, analysis and visualisation features. Nevertheless, its functionality is
of general benefit to QGIS users by introducing tools for exploratory spatial data analysis. In this
article we present the Space Syntax Toolkit, describing its principles and implementation, the
features of its two initial modules Graph Analysis and Attribute Explorer and give indication
of future developments.

Keywords
GIS, QGIS, space syntax, network analysis, urban design education.

J Gil, T Varoudis, K Karimi & A Penn 148:1

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

1. Introduction
Urban design practice, as a discipline closely related to architecture, has traditionally relied on the
use of Computer Aided Design (CAD) platforms for computational tasks. However, urban design
processes in evidence-based design, comprehensive planning and sustainable urban development
have been evolving to require increasingly complex decision making in many fields, e.g. social,
economic, environmental, transportation. These advanced design processes rely on a range of
support tools, including spatial and statistical analysis, of which space syntax is one of the frequently
used theories and methods. Over the years, a range of software has been developed to perform
1
space syntax analysis , as standalone tools (e.g. Axman and WebMap by Nick Dalton, Depthmap
(Turner, 2007), Mindwalk by Lucas Figueiredo), as extensions to CAD (e.g. Spatialist from Georgia
Tech, sDNA (Chiaradia et al., 2012), UNA Toolbox (Sevtsuk and Mekonnen, 2012)) or to Geographic
Information Systems (GIS) (e.g. Axwoman by Bin Jiang, UNA Toolbox, sDNA, Confeego and iVALUL
from Space Syntax Ltd). GIS seems to offer a natural platform for such methods (Jiang and
Claramunt, 2002; Gil et al., 2007; Chiaradia et al., 2008), with its ability to handle geographic and
geometric data associated with attribute information, to easily perform spatial, mathematical and
statistical calculations, and to visualise the results. In order to improve space syntaxs contribution to
urban design research, education and practice, it is important to develop support tools that integrate
its complete workflows in a single platform, are intuitive for beginners, flexible for researchers,
operational for practitioners, and accessible to all users. With this aim we have created an open
source software project to develop a GIS toolkit that integrates the depthmapX multi-platform
spatial network analysis software (Varoudis, 2012) with the QGIS software platform (QGIS
Development Team, 2014). The Space Syntax Toolkit (SST) (Gil, 2014) provides a frontend for
depthmapX within QGIS. It enhances space syntax workflows with GIS data, analysis and visualisation
features, integrates space syntax network analysis in QGIS, and expands QGIS with new exploratory
data analysis functionality. In this article we present the SST, describe its implementation
requirements and principles, the features of its two initial modules Graph Analysis and Attribute
Explorer and indicate future developments.

2. Requirements for and development of a toolkit

Integrated workflows for QGIS

The SSTs aim is to provide an entry into the use of GIS for advanced spatial network analysis
methods, primarily in post-graduate education and research. The key challenges are related with the
engagement of users that have a visual and architectural background in the use of GIS, avoiding as
much as possible the frustrations inherent to complex analysis workflows, with complicated and
lengthy intermediate steps that require specific input/output formats, and instead offering swift and
intuitive data exploration. For space syntax beginners, the SST should offer an interface in a single
platform to support complete analysis workflows, with a clear user interface using space syntax
specific terminology and offering simple options. For more experienced researchers and
practitioners, the SST should allow advanced analytic options and support greater exploratory depth
of the results. Along with these user experience requirements, the tool should be flexible in dealing
with a variety of data sets, supporting large network models and streamlining the workflows for
greater speed, eliminating as much as possible error prone and cumbersome intermediate steps.
Figure 1 illustrates (a) the current space syntax analysis workflow in depthmapX (Varoudis, 2012) for
CAD and GIS users, versus (b) the desired workflow integrated in a single and expandable GIS
platform. The most obvious advantage is the elimination of several data export steps with an
accumulated benefit, as each data export entails the selection of file formats, versions, projections
and other settings for both the axial map and the unlinks, which set wrongly will cause problems in
the analysis or the interpretation of results. Furthermore, these steps have to be repeated every
time the user needs to make changes to the axial model, for example to correct a mistake in the axial
model. In the desired workflow, the user should be able to carry out what could be considered
standard space syntax analysis practice without the use of additional software applications.

1
https://en.wikipedia.org/wiki/Spatial_network_analysis_software

J Gil, T Varoudis, K Karimi & A Penn 148:2

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

Figure 1: Diagrams of the current (a) and desired (b) workflow of space syntax analysis using the Depthmap or
depthmapX softwares.

The typical space syntax research workflow goes through a series of stages that define the user
requirements for a comprehensive SST. Table 1 offers a list of such stages and operations adequate
for urban research, including data acquisition, geo-processing of spatial data sources, network model
preparation and analysis, statistical exploration and analysis, and production of maps and charts for
reporting. This list is necessarily incomplete, as much depends on the specific research topic,
questions and methodology.

J Gil, T Varoudis, K Karimi & A Penn 148:3

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

Table 1: Stages and operations of a typical space syntax urban research workflow.

Stage Operations

Data acquisition Collect base map data

Convert data formats
Georeference vector and raster data
User behaviour data entry
Location survey data entry

Geo-processing Filter, aggregate and join data based on attributes and spatial relations
Find nearest feature
Transform geometries
Attributes update and calculation

Network model preparation Draw map

Extract map from larger data set
Prepare plans
Generalise road network
Map verification
Manage unlinks and links

Network analysis Analyse network models (space syntax)

Post process analysis results (attribute naming, selection, new calculations)
Calculate shortest route and catchment areas
Street and block morphology

Exploratory spatial data analysis Interactive visual exploration of results

Inquiry of individual feature values
Identify the core of the network
Spatial analysis (interpolation, hotspots, clustering)
Terrain analysis

Statistical analysis Descriptive statistics

Transform columns
Interactive charts (histogram, scatter plot)
Linear regression
Multivariate regression
Statistical clustering (PCA, k-means, hierarchical)
Export data to statistics packages

Reporting and visualisation Create base map compositions for print

Set output format and resolution
Standard visualisation (colour ranges) for screen and print
Store named views
Produce map series based on attributes

Implementation in QGIS
The SST is implemented in a Python plug-in for the QGIS platform, which offers various advantages
as a project and software to the space syntax community. The project is an open source project
member of the OSGeo Foundation, meaning that it is compliant with open standards, incorporates
many OSGeo analysis packages, and is free to use for any purpose. Furthermore, it has a growing
support and user community with an active development cycle and a growing number of plug-ins
developed by this community. The software itself is cross platform running on Windows, Linux, Mac
OSX and Android, with a user friendly interface emulating in many respects that of ArcGIS, the
leading commercial GIS platform, and is documented in English and other languages. It supports
most input and output data formats (i.e. raster, vector, databases, web services, tabular data), offers
managed plug-in installation, supports automation with macros and scripts, and more advanced
customisation with its Python API, that gives access to useful packages developed by the wider
Python community.

J Gil, T Varoudis, K Karimi & A Penn 148:4

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

QGIS can perform, either natively or using existing plug-ins, many of the user requirements outlined
in Table 1 that are of a general GIS nature. Nevertheless, to meet those requirements, QGIS still
presents several challenges when compared to other platforms currently used by space syntax
researchers. On the one hand, it lacks some advanced spatial analysis and statistics available in
ArcGIS, and when these are available their use is not intuitive, can be slow and unreliable, and
invariably the results are in new tables that require additional (elaborate) operations to merge. On
the other hand, it lacks support for exploratory analysis available in depthmapX, such as interaction
between statistics and the map canvas, and a fast preview for the long lists of numeric attributes
that result from the analysis. In addition, it does not currently have plugins to support space syntax
specific workflows, such as axial map production, unlinks management, survey data entry, or
network centrality analysis. The aim of the SST is to cover some of these gaps, in particular those
related to the integration of depthmapX analysis and the implementation of its user interface
features.

The toolkits architecture

The architecture of the SST consists of a single plug-in for QGIS containing separate modules
developed similar to independent plug-ins to allow easier expansion of the toolkit, and a set of
shared utility functions to increase consistency between modules and facilitate development of new
features. Each module of the SST is accessible individually from the menu and toolbar developed as a
side panel (modeless window) allowing the continued use of QGISs other features. Rather than
performing single shot operations from multiple dialog boxes, these modules aim to function
alongside and in support of the users standard workflow, integrating the specific features and tools
required. Currently there are two modules, Graph Analysis and Attributes Explorer that are
presented in detail in the next section of the paper.

The SST has dependencies on other packages that are not present in QGIS by default. Of special
importance is the integration of depthmapX in the Graph Analysis module to perform the standard
space syntax network analysis calculations. This integration uses, a newly developed, special version
of depthmapX called depthmapX[net] that was developed by Tasos Varoudis in order to deliver all
internal analysis capabilities through simple network calls to external applications or scripting
environments (e.g Grasshoper plug-in; JAVA and Python based depthmapX client by Tasos Varoudis).
The only requirement is that depthmapX[net] is installed and running when performing a space
syntax analysis, and the Graph Analysis module works as a remote control, sending the data and
analysis parameters and retrieving the results when the calculation is completed. Additional Python
packages are required to use the full features of the SST, namely igraph (Csardi and Nepusz, 2006) to
do alternative network analysis, and pyqtgraph (www.pyqtgraph.org) for displaying interactive
charts. If these packages are missing from the Python installation, the user will still be able to use the
SST, albeit with some limitations, as the alternative features are not completely equivalent.

The SST has been developed to work with any vector layer loaded in a QGIS map canvas, and
currently supports output to the native data formats of QGIS: shape files, spatialite personal
geodatabases, PostGIS geodatabases and memory layers. Each of these data formats has unique
characteristics and the SST takes advantage of format specific features, such as the spatialite and
PostGIS SQL query engines to optimise performance in many operations, particularly noticeable with
larger data sets. This is a new approach in QGIS plugins, which by default use the shape file format.
For this reason users should become familiar with the pros and cons of each data format, and choose
the most suitable to their needs.

3. The Space Syntax Toolkit plug-in

The SST plug-in is available for installation and update via the Plugins Manager tool of QGIS, making
it easily accessible and maintainable to any QGIS user. Currently, the SST plug-in consists of two
modules, Graph analysis and Attributes explorer, whose features are described next.

J Gil, T Varoudis, K Karimi & A Penn 148:5

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

Graph Analysis module features

The Graph analysis module supports the verification and analysis of the space syntax network
model, representing the public open space network. It works with the axial line and segment models,
including a unlinks layer to indicate bridges and tunnels where there is no level crossing between
lines. The layers that make up the model are selected in the respective tabs. The module has a
verification tool to check the geometric and topological integrity of each layer, helping to correct
problems before running the analysis. The axial and segment analysis is performed in
depthmapX[net] with the remote link from QGIS.

Layer verification
The layer verification tools in the Verify layer tab check if the model layers are correct and have the
characteristics required for analysis (Figure 2). The axial map verification checks for geometry and
topology mistakes that can easily occur when digitising the axial lines over digital cartography, or
when merging maps from different authors. The mistakes checked include:

Small line lines with a length below the minimum (default: > 1 m)
Polyline lines made up of more than two nodes
Coinciding points lines with two points coinciding
Duplicate geometry two lines with identical geometry
Short line line whose endpoint is close to another line without intersecting it (default: > 1
m)
Orphan line that is not intersecting any other line
Island group of lines that are disconnected from the main map (only verified if the igraph
package is present)

Figure 2: Graph Analysis panel with axial map verification report.

J Gil, T Varoudis, K Karimi & A Penn 148:6

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

The distance thresholds used in the verification process can be adjusted if required via a separate
settings window, to keep the user interface as simple as possible. When the verification process
concludes, any problems found are listed in the report window and the total number of occurrences
of each problem type is indicated in a menu. The verification tool supports the user to manually
correct any problems identified by filtering the list of problems by type, and by automatically
selecting and zooming into the line or group of lines corresponding to an entry in the report. The
user can then use the QGIS layer editing features to modify or remove the line(s).

When verifying the axial map, if the Unlinks layer is selected, these are taken into account and can
reveal additional topological problems due to unlink disconnections. However, this unlinks layer
must also be verified and corrected for the axial map verification results to be accurate. An unlink
can be any map feature that identifies an intersection between two axial lines to be disconnected,
i.e. it can be a point near the intersection, or a line or polygon intersecting the two relevant axial
lines. The unlinks layer verification checks for topological and id problems in relation to the axial
lines layer:

Multiple lines unlinks intersecting more than two axial lines

Single line unlinks intersecting only one axial line
No lines isolated unlinks not intersecting any lines
Same line id unlinks with the same line id attributes
Unmatched line id unlinks with the line id attributes not matching the currently
intersected axial lines

Figure 3: Unlinks verification reports, before (a) and after (b) updating the line ids.

The first time the unlinks verification tool runs, it checks if two attributes exist (i.e. line1 and line2)
that identify the lines to be unlinked. If these attributes are missing, the problems report includes all
unlinks (Figure 3a) and the user must run the Update IDs tool assigns the id of the lines at the
intersection identified by each unlink. This update needs to be triggered by the user to prevent
(incorrect) automatic id changes after axial map edits. In such cases, any unlinks with unmatched line
id will be signalled and can be confirmed by the user, only then updating the line ids. The unlinks
problem selection, filtering and editing process is identical to that for axial maps.

J Gil, T Varoudis, K Karimi & A Penn 148:7

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

depthmapX remote
The depthmapX remote tool controls the graph analysis of the selected map and unlinks layers,
with the same settings that one would find in depthmapX but in a more compact layout (Figure 4),
with advanced settings available from a separate settings window. One can choose between axial
(topological) and segment (angular) analysis, set the radius distance, and select an attribute to use
for weighted analysis, and select the output layer name.

Figure 4: depthmapX remote tab (a) for setting up and running axial and segment analyses remotely, with
Advanced settings (b) for changing some of the default analysis parameters.

The layers data and analysis parameters are sent to depthmapX[net], and both QGIS and
depthmapX[net] must be kept running throughout the process, although one can continue working
on both while the analysis runs in the background. Once the analysis is completed, the post-
processing stage takes place on the QGIS side, where the results are filtered to select a relevant set
of measures, new standard measures are calculated, and the attributes are renamed to be
compatible with the output format, e.g. shape files only support attribute names with up to 10
characters. The results layer is then added to the current map canvas and the new attributes can be
seen in the attributes table. However, the best way to analyse the results is to use the Attributes
Explorer module.

Attributes Explorer module features

The Attributes explorer module supports the visual and statistical exploration of network analysis
results, with the aim to emulate the interactive nature of depthmapX. It offers the quick analysis of
individual quantitative attributes of a selected layer, displaying the values on the map using
simplified symbology, providing descriptive statistics, and plotting basic interactive charts (Figure 5).

J Gil, T Varoudis, K Karimi & A Penn 148:8

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

Figure 5: QGIS window with the Attributes Explorer panel docked on the right.

Attributes symbology
The Symbology tool (Figure 6) supports the display and production of maps that are standard within
the space syntax community and suitable for academic publication. The default style is applied as
soon as an attribute is selected in the Numeric attributes list, but the user can change the colour
and data range settings to create a custom symbology that is more appropriate to the specific
attribute. The symbology settings of the different attributes are stored in the QGIS project and
become the style associated with those attributes when using this module, allowing the alternate
visualisation of different attributes based on their individual styles.

Figure 6: Attribute symbology settings tab.

J Gil, T Varoudis, K Karimi & A Penn 148:9

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

The symbology settings cover the display style and data range. The display style sets the colour scale
(i.e. space syntax classic, red-blue, grey scale and monochrome, shown in Figure 7 a through c), with
options to invert the scale and sort the colours, ensuring that the top or bottom values are displayed
first on the map, and set the line width. The grey scale and monochrome colour ranges use black or
white lines depending on the backgrounds colour (Figure 7d). The data range sets the number and
type of colour intervals, e.g. equal range, quantiles, natural breaks or custom. The custom interval is
by default the maximum (top) and minimum (bottom) values of the attribute, but the user can
change the size of the top and bottom range, using a percentage or the absolute value. The other
intervals in between are automatically calculated using equal ranges distribution. This setting is
equivalent to the slider controls in the Colour Range settings of depthmapX.

Figure 7: The different colour range settings: classic (a), red-blue (b), greyscale (c) and monochrome (d) with
varying line width. The greyscale and monochrome colour ranges automatically use black or white lines
depending on the background colour.

Attributes statistics and charts

The Statistics tool (Figure 8) displays a table with descriptive statistics of the selected numeric
attribute, and of the currently selected features from the same layer. These statistics give an

J Gil, T Varoudis, K Karimi & A Penn 148:10

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

overview of the overall distribution and range of values, and are helpful to understanding the
results, identifying anomalies, and decide on the most appropriate symbology settings. The Charts
tool (Figure 9) shows simple interactive charts of the selected attribute. The histogram and scatter
plot can be zoomed in and panned freely by the user, and the currently selected features are
displayed in red. For the scatter plot (Figure 9b), the Y axis (dependent variable) remains constant
allowing the quick comparison of different independent variables selected in the attributes list.
These charts are only available if the pyqtgraph Python package is installed.

Figure 8: Descriptive statistics of the selected attribute values, with the statistics for the data of the selection on
the map.

Figure 9: Histogram (a) and scatter plot (b) of the selected attribute values. In red are the data points
corresponding to the selected features on the map.

4. Future development
The SST is a long-term open source project currently managed by the Space Syntax Laboratory at the
Bartlett School of Architecture, UCL. The current release only covers a fraction of the stages and
operations of space syntax research workflows listed in Table 1. While some of the more general
functionality will probably be added to the fast evolving QGIS project, the SST should continue to
grow, with fixes and enhancements to existing functionality, new functionality added to the current
modules, and the development of new modules. Some improvements planned for future versions of
the SST, namely: adding two way interaction between the map and the attributes explorer charts; a
catchment area analysis module; or an urban morphology module for street and block metrics.

J Gil, T Varoudis, K Karimi & A Penn 148:11

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS
SSS10 Proceedings of the 10th International Space Syntax Symposium

As an open source project, open to the wider communities of space syntax researchers and QGIS
users, the direction of development should be informed by the needs of that community, and open
to contributions in the form of new code, fixes, testing or documentation.

5. Conclusion
In this article we introduced the Space Syntax Toolkit, a Python plug-in for QGIS that integrates
depthmapX[net]. The use of GIS with space syntax analysis is of growing importance to address a
wide range of empirical studies. However, the space syntax research process has a set of specific
stages and operations that can benefit from integrated workflows in GIS, making it more accessible
but also efficient to a range of different users. The SST implements with its initial modules the basic
workflow of axial and segment map analysis, from map production to publication of the results,
emulating in part the exploratory spatial analysis features of depthmapX. This should offer a
smoother entry into QGIS to space syntax beginners but also to existing depthmapX users who do
not use GIS. This open source project will continue to evolve, to offer a broader set of features to,
and with the support of, the space syntax research and practice community.

Acknowledgements
This work has been funded by Platform Grant Extension: Space, Time and Interaction EPSRC
GR/S64561/01.

We would like to thank the UCL researchers that offered insight into their practices to help define
the user requirements. In particular, we would like to thank Stephen Law and colleagues at Space
Syntax Ltd for their detailed analysis of the workflows and stages of space syntax projects. Finally, we
would like to thank the early adopters of the SST for their feedback and questions.

References
Chiaradia, A., Schwander, C., Gil, J., Friedrich, E., Gosset, A., (2008). Mapping the intangible value of urban
layout (i-VALUL): Developing a tool kit for the socio-economic valuation of urban areas, for designers
and decision makers, In: Timmermans, H., DeVries, B. (Eds.), 9th International Conference on Design &
Decision Support Systems in Architecture and Urban Planning. Eindhoven University of Technology,
Eindhoven, The Netherlands.
Chiaradia, A., Webster, C., Cooper, C., (2012). sDNA - A software for spatial design network analysis. Cardiff
University, Cardiff, UK.
Csardi, G., Nepusz, T., (2006). The igraph Software Package for Complex Network Research. In InterJournal
Complex Systems, p.1695.
Gil, J., (2014), Space Syntax Toolkit for QGIS, Version 0.1.0, Open Source, Available at:
https://github.com/SpaceGroupUCL/qgisSpaceSyntaxToolkit
Gil, J., Stutz, C., Chiaradia, A., (2007). Confeego: Tool Set for Spatial Configuration Studies, In: Turner, A. (Ed.),
New Developments in Space Syntax Software. Presented at the 6th International Space Syntax
Symposium, Istanbul Technical University, Istanbul, p.1522.
Jiang, B., Claramunt, C., (2002). Integration of Space Syntax into GIS: New Perspectives for Urban Morphology.
In Transactions in GIS 6, p.295309. doi:10.1111/1467-9671.00112
QGIS Development Team, (2014). QGIS Geographic Information System. Open Source Geospatial Foundation
Project. Available at: http://qgis.osgeo.org/en/site/
Sevtsuk, A., Mekonnen, M., (2012). Urban network analysis. A new toolbox for ArcGIS. In Revue Internationale
de Gomatique 22, p.287305.
Turner, A., (2007). UCL Depthmap 7: From Isovist Analysis to Generic Spatial Network Analysis, In: Turner, A.
(Ed.), New Developments in Space Syntax Software. Presented at the 6th International Space Syntax
Symposium, Istanbul Technical University, Istanbul, p.4351.
Varoudis T., (2012), 'depthmapX Multi-Platform Spatial Network Analysis Software, Version 0.30 Open Source.
Avaialble at: http://varoudis.github.io/depthmapX/

J Gil, T Varoudis, K Karimi & A Penn 148:12

The space syntax toolkit: Integrating depthmapX and exploratory spatial analysis workflows in QGIS