SSS8 - Chile 2012Time use and movement behaviour of young

people in citiesThe application of GPS tracking in tracing

movement pattern of young people for a week in Aalborg

The Aalborg Case• Thank you very much for the possibility to participate in the

SSS8

• Thank you for the critique to our paper/presentation• Thank you for placing the presentation in this session

• A special thank you to Akkelies Van Ness, Delft University of Technology, Who introduced us to the magnifying world of Space Syntax Analyses• And to Margarita Greene Z. and José Reyes S. from the

SSS8 Organizing Committee for answering our e-mails

Diverse Urban Spaces

• Our research profile• Research group located at Aalborg University’s

Department of Architecture, Design and Media Technology in Denmark 6 members

• Employees with various academic backgrounds, ph.d.’s, post doc., architects, surveyors, sociologists, student assistants etc.

• Our research work• Studies of mobility among different population groups

across varying scales in both indoor and outdoor environments

• Examples of our research• Analyses of how citizens/humans in an urban

environment make use of over parks / plazas / central city areas e.g. – GIS/GPS mapping of the everyday movement and time consumption patterns of 300 high school students (2007), 200 Bicyclists (2011), and 400 individual respondents in around 100 families (2011)

Motivation for applying Space Syntax on GPS data

• All our projects – especially our flagship project ”Diverse Urban Spaces” – result in very rich GPS data sources describing actual movement and behaviour, compare to others types of datasamples

• Akkelies Van Ness, Delft University of Technology, provided a series of Space Syntax analyses of Aalborg City’s road and path network

• Two data sources depicting the infrastructure (the space syntax data and the GPS data)

• Comparison in order to evaluate theory versus practice and as such the quality of the Space Syntax method

• Ratti, C. 2004b, "Space syntax: some inconsistencies", Environment and Planning B: Planning and Design, vol. 31, pp. 487-499.

Diverse Urban Spaces in a nutshell

• A grand research project involving over 300 young people attending high school or equevalent level of education in Aalborg Denmark who were tracked in 7 days

• Research objective: To study how the city og Aalborg is used by this particular segment of respondents

• Technical setup was twofold:• Each respondent had to carry a hand-held

GPS-receiver throughout a week• After each survey day, the respondent was

tasked with filling out a trip diary

• Motivation for continuous participation wasdaily and weekly lotteries with cash prizes

Technical setup

Case Data cleansingAnd preparation

Server / databaseData gathering GIS analyses

• In the Aalborg Case we used a subsample of data from 169 statistical verified respondents tracked 24 hours in 7 days from the 300 respondents

Example of output analysis - Accumulated time consumption

Central city area in Aalborg, Denmark

Urban life ??

Overall - Space syntax analyses

• Conducted by Akkelies Van Ness, Delft University of Technology

• Different analyses using varying settings

• High metrical radius highlights main transportation corridors

• Low metrical radius highlights local city centres

• Output stems with reality (with exceptions) for Aalborg, cf. next slide• Main transportation corridors Vesterbro, Hobrovej,

Sønderbro, Østre Allé get highlighted• Pedestrian areas in Aalborg and Nørresundby get

highlighted• Some noise in both analyses

Overall - Space syntax analyses

Global integration – high metrical radius

Map uses in the analyses

Local integration – low metrical radius

Time consumption of men.

Time consumption of woman.

Put the two together

• The analyses depicting the global integration was chosen for the comparison

• The reason for using this analysis is that the dataset containing mobility and time consumption involves the entire urban area

• Time accumulation Maps - The largest time was spent in the various shopping areas in Aalborg in the central city area. Secondly, some high amount of time was spent on the various main routes leading through and between urban areas.

• Woman spent more time in central shopping areas

Overall - Comparison method

• Founded on an assumption of positive linear correlation between a road segment’s rank as classified by the space syntax analysis and the accumulated time consumption on streets created by the respondents in the Diverse Urban Spaces project

• ”Home-spun” procedure involving 3 steps

Comparison method, step 1

• 10 categories of time consumption levels are calculated, corresponding to the 10 space syntax classification levels.

• Time consumption categories are yielded as quantile values based on the time consumption data

Overall - Comparison method, step 2

• Each observation of accumulated time which spatially intersects a given road segment is selected

• A mean value of accumulated time consumption for the road segment is calculated based on the values of the selected observations

Comparison method, step 3

• The mean value (µ) is evaluated against the quantile values corresponding to the the space syntax classification

• If µ matches the quantile values, coherence is attained

Space syntax class

Coherence at (strict rule)

Coherence at (loose rule)

10 µ < q1 µ < q2

9 q1 < µ < q2 µ < q3

8 q2 < µ < q3 q1 < µ < q4

7 q3 < µ < q4 q2 < µ < q5

6 q4 < µ < q5 q3 < µ < q6

5 q5 < µ < q6 q4 < µ < q7

4 q6 < µ < q7 q5 < µ < q8

3 q7 < µ < q8 q6 < µ < q9

2 q8 < µ < q9 q7 < µ1 q9 < µ q8 < µ

Overall Men/Women - Results, strict coherence rule

Men Women

Results, strict coherence rule

• Generally only coherence along the main transportation corridors

• There is coherence between time spent and space syntax classification, when the mean value of the selected squares is compared with the expected time consumption for the polyline based on the classification. The expected value is derived by yielding 9 quantile values (q) which divide the time consumption registration dataset into 10 approximately evenly distributed groups.

Results, loose coherence rule

Men Women

Results, loose coherence rule

• A greater degree of coherence in suburban areas as well as some parts of the city centres in addition to the main transportation corridors

• A close-up view of the main pedestrian areas shows lack of coherence between space syntax classification and time consumption

• The coherence rule is loosened slightly in the sense that coherence is achieved if the mean value resides within a buffer of ± 1 quantile of the expected time consumption value. I.e., coherence is reached for the space syntax class 5 at q4 < µ < q7 instead of q5 < µ < q6.

Still lack of coherence in the main pedestrian in the central City areas for men and woman

Results, overall statistics

  Occurrences of coherence 

Occurrences of lack of coherence 

Coherence-ratio 

Women, strict rule

637 4,429 10.86 %

Men, strict rule

625 4,461 10.65 %

Women, loose rule

1,972 3,094 33.62 %

Men, loose rule

1,916 3,170 32.67 %

Number of road segments 5865

  Women MenNon-visited road segments

799 779

Number of squares 14,891 15,625Accumulated time consumption

15,125,711 seconds 18,838,688 seconds

Evaluation and conclusions• High school students in Aalborg also travel frequently along

main - transportation corridors

• There is no coherence between space syntax classification and actual time spendure in the main pedestrian areas in the central City area.Most likely because the used Space Syntax analysis in this cases uses a high metrical radius which doesn’t highlight road segments with a high local integration but this will change when using a more loose coherence rule and maybe low metrical radius in the Space Syntax analysis.

• The lack of coherence is a natural consequence when the space syntax analysis is executed with a high metrical radius. As such, the northernmost shopping districts have a classification which is too low to attain coherence with the massive time consumption registered in these areas. If the space syntax classification is mapped on top of the time consumption dataset, this assumption becomes more reliable.

What can be done to adjust Space Syntax to GPS data• One of the points of criticism of the space syntax method

was a lack of the metrical properties in their analyses (Ratti 2004). Now it is incorporated in the calculations. As research results show, the geometrical and topological distances correspond with the pedestrian and vehicle flow rates and the location pattern of shops more than the metrical distances. However, when applying metrical radiuses in the angular and axial analyses, some striking results can be seen.

• Streets with high integration values with a high metrical radius tend to be the potential routes for through movement. Conversely, streets with high integration values with a low metrical radius tend to be potential meeting places for the neighbourhood. When comparing these two analyses with one another, the most vital urban areas tend to be where streets have high integration values with both high and low metrical radiuses. (van Nes, Berghauser-Pont, Maschoodi, 2011).

Further work• Conduct the comparison

using a better scientific founded method such as Linear Regression instead of the “homespun” method

• Conduct the comparison for centre areas using the Space Syntax analysis with a low metrical radius – will most likely lead to a higher comparison rate

• Conduct the comparison based on amount of trips and not accumulated time

• Define what urban life is ..

Thank you for your attention

Any questions?