cityware: design and evaluation of city-scale pervasive technologies tim jay [email protected]
TRANSCRIPT
Contents
Background and Rationale Methodology Example studies and findings
The context logger Perceptions of the environment: neighbourhood
and city Bluefish: A privacy probe
Discussion
Cityware
Research collaboration between the University of Bath, Imperial College, The Bartlett at University College London, Vodafone, Nokia and Hewlett-Packard (http://www.cityware.org.uk)
Aim: To increase our knowledge and understanding of people’s relationships with urban spaces and with pervasive technologies in order to enable the development of tools and techniques for the implementation of long-term, city-scale pervasive systems.
How do people understand the city? Lynch, 1960 – The Image of the City
Imagability/legibility Basic elements – paths, edges, districts, nodes
and landmarks Wayfinding and Navigation Landmarks Cognitive Maps Urban Semiotics
The relationship between pervasive technologies and urban environments
Urbantapestries.net, Urban-atmospheres.net, MobileBristol.co.uk (all worth having a look at)
Often cover relatively small physical areas Often run over short timescales Cityware aims to address the methodological
challenge of longitudinal, city-scale studies
Some issues with traditional mobile and pervasive evaluation Bias towards laboratory testing Sample bias – technologically literate
participants Relatively restricted in both time and space –
often out of context
The Cityware Cohort
30 participants, working with us for the 3-year duration of the project
Aged 14-80+ 20 male, 10 female Variety of occupations and levels of technological
ability/awareness All equipped with a Nokia Smartphone and
Vodafone SIM Attend regular workshops + complete other activities
Data Collected
Questionnaires, interviews & focus group work
Personality data Map sketching Moblogging
Example Studies
The context logger Perceptions of the environment –
neighbourhood and city Bluefish: A privacy probe
The context logger
Developed in the Computer Science department, Imperial College
Designed to inform the development of context-aware applications and services
Records a variety of data generated by a mobile phone
Data is automatically transmitted to a remote server using the 3G/GPRS network
Context logged
Time/date and sender/receiver of text messages Time/date and sender receiver of voice calls Camera images Applications in operation Location (using either Bluetooth GPS unit, or built-in
functionality) Bluetooth scans Cell ID in use Active profile Battery strength
Context logger evaluation
Iterative process Local group 3 cohort members (selection criteria) Full cohort
Refinement of the application at each stage Functionality Interaction with general use of phone Privacy and trust concerns
Context Logger: Discussion
Different issues arise at various points of iteration
There is an association between the demographic of the sample and the level of concern that arises
Some issues could only have been discovered through genuine use with genuine user group
Perceptions of the environment How do city residents and visitors perceive and use
the city? Some background
Space syntax methods used to predict pedestrian movements. These can be compared with:
Bluetooth gate counts used in order to analyse actual movements of Bluetooth device around the city (+ social network analysis)
These don’t give access to some important aspects that we are interested in – what meanings are derived from, or ascribed to, the city?
Methods
3 phases Neighbourhoods City ‘Real’ City
Long-term study with cohort Multiple methods
Methods
Moblogging tour Questionnaires
Landmarks City use
Map sketching Boundary marking
The Neighbourhood/City Moblogging Tour GPS trail Photos Video Text notes Audio notes Analysis focuses on:
Representations across formats Landmarks Effects of moblogging on perception
Differences in representation
Landmarks
Rural v. Urban Traditional v. Personal Traditional landmarks
Churches, schools, war memorials, parks, pubs etc.
Personal landmarks (and non-landmarks) Allotments down a back lane A derelict building
Effects of moblogging on perception Change in scale between 1st and 2nd sketch
maps
Next steps
Recently completed similar set of activities in the city centre
Day-to-day city use Expand dataset – including digital landmarks,
Bluetooth, Wifi
Bluefish
Designed as a ‘privacy probe’ Previous Cityware research has shown that many people
are not aware that they are broadcasting information about themselves via their Bluetooth devices
Consists of a network of Bluetooth scanners + 3-4 public display screens, in various locations across Bath
The screens represent Bluetooth activity that has been recorded by the network of scanners - Example
Research Questions
How do people respond to their data and history being made visible by the Bluefish system? How does behaviour change – e.g. Bluetooth
turned on/off, opting in/out of the system What are people’s perceptions of the data, and
what are the reasons for any behaviour change?
Methods
System logs can be used to analyse changes in behaviour but don’t give us the reasons
Interviews on location can tell us something about people’s perception of the system and any reasons for changes in behaviour But – some of the more interesting data will not be
accessible through interviews Initial work will be done with the cohort
Advantages of the cohort
Broad demographic Trust and investment in the project Gives us access to participants who may
have concerns over privacy/security/trust issues, or who are not engaged with the system
Summary
Sustained engagement of participants Helps address the problem of validity
Combinations of data Mixture of methods – lab + field, quantitative +
qualitative Engagement of a community
Participants have ownership of and commitment to the project