Workshop B4. The Collection and Processing of Survey Data Using Mobile Technologies.

Workshop B4. The Collection and Processing

of Survey Data Using Mobile Technologies Workshop B4. The Collection and Processing

of Survey Data Using Mobile Technologies

Workshop Participants

Workshop Chair: Jean Wolf, United StatesResource Paper Author: Peter Stopher, AustraliaDiscussant: Barbara Noble, United KingdomRapporteur: Sean Doherty, Canada

Contributing AuthorsStephan Krygsman, South AfricaMaat Kees,The NetherlandsNadine Schüssler, Switzerland

Other ParticipantsTracy Anderson Robert Follmer Roger Mackett Juan de Dios Ortuzar

Patrick Bonnel Martin Kagerbauer Steve MacFeely Matthew Roorda

Andrew Clarke Nina Karasmaa Fabrice Marchal Daniel Sauter

Kelly Clifton Thipphaphone Khenmingmongkhon

Philippe Marchal Yuan ShuningSusan Swain

Eric Cornelis Nick Lanyon Michael Meschik Liva Vagane

Notes:

• Sizable group (n=28)

• High attendance and participation rates

• No incentives offered

Workshop Scope

Really did focus on data collection and processing using mobile technologies

Tried to stay away from non-response Limited discussion on broader study design

State of the Art - Devices & Data Collection

Device evolution is heading to wearables, but vehicle-based studies may still have a role

Most studies include GPS and diary, but trend towards GPS only Interview and delivery deployment methods vary Viable alternatives to GPS are being evaluated

• Mobile/cell phone options (now) Also used for CAPI• Bluetooth / WiFi / RFID / smartcard / RDS (in research)

Alternatives have different levels of detail, precision, cost, coverage

Mixed modes (tech and non-tech) could be used Select methods and solution(s) based on purpose / need

Various algorithms developed to identify key diary elements • stops, trips – OD / route / distance / start time / duration• travel mode• trip purpose

Use of GIS datasets important (road network, transit network, points of interest, land use, etc.)

Still largely in exploratory phase • Some implementing rule based, others using fuzzy logic

Little validation and little ‘ground truth’ to do it Prompted recall interfaces developed and implemented, but

burden is an issue Mobile/cell phone research showing promise

State of the Art – Data Processing

Deploying/retaining devices On/off switch desirable ? Logging rules (e.g., frequency, speed screen) Missing data / messy data Age cut off for deployment (practicality and ethics) Reducing bias Encryption/security Ethics Avoiding lawsuits Retaining/archiving raw data Cost of cell-based location for large samples Practicality of large scale location-enabled mobile phones

(commercial considerations) How best to recruit and communicate with subjects

Issues - Devices and Data Collection

Lack of available software – commercial, share/free ware Are we ready to standardize?

• Data Collection (NMEA parsing)• Data Storage (XML desirable)• Algorithms

Calibration/validation data essential, but challenging to acquire• Prompted recall offers some potential• Direct observation may be needed

Assessing respondent burden from prompted recall Different methods used, e.g., fuzzy logic, AI, rule-based,…

Issues - Data Processing

Research Needs – Data Collection

Continued improvements in mobile technology devicesContinued improvements in mobile technology devices• Functionality, cost, power capacity, storage capacity, etc.Functionality, cost, power capacity, storage capacity, etc.

Deployment method analyses with respect to study purpose, cost, Deployment method analyses with respect to study purpose, cost, response ratesresponse rates

Continued research on mobile phone ‘tower location’ traces for Continued research on mobile phone ‘tower location’ traces for travel surveystravel surveys

Feasibility study for deployment of custom software for Feasibility study for deployment of custom software for population-based mobile phone user sample population-based mobile phone user sample

Possible joint venture research – government, private, university Possible joint venture research – government, private, university (e.g., health or tourism research)(e.g., health or tourism research)

How young should or could participants be?How young should or could participants be? Impact of Galileo on accuracy and coverage?Impact of Galileo on accuracy and coverage?

Research Needs – Data Processing

Standard data, standard processing software, or perhaps algorithm Standard data, standard processing software, or perhaps algorithm modulesmodules

Commercial or free processing software (are we ready?)Commercial or free processing software (are we ready?) Comparative analysis of accuracy of mode and purpose identificationComparative analysis of accuracy of mode and purpose identification Development of travel companion (who) / party size estimators (and Development of travel companion (who) / party size estimators (and

is this really needed?)is this really needed?) Development of validation datasetsDevelopment of validation datasets Development of algorithms / software independent of GISDevelopment of algorithms / software independent of GIS Overall methodology (GPS/CATI, GPS/PR, GPS only) tradeoff analysis Overall methodology (GPS/CATI, GPS/PR, GPS only) tradeoff analysis

(burden, quality, bias, cost). Same for mobile phones.(burden, quality, bias, cost). Same for mobile phones. Pushing the modelling paradigms (e.g., number of travel days, Pushing the modelling paradigms (e.g., number of travel days,

number of persons per household, data elements, travel time number of persons per household, data elements, travel time measurements) measurements)