Solution Tag: Innovation

Development of a geolocation plugin

Posted on February 6, 2020 by -

Context

Eurostat started a project to modernize the data collection by developing innovative and shareable solutions among countries and to contribute to the European platform for Trusted Smart Surveys.

Time Use Surveys are an excellent example to show how sensor data can improve the data quality and collection efficiency.

This trajectory starts with the inclusion of sensor-derived geolocation as example.

State of play

Time Use Surveys have a long standing tradition and are recognized by international organisations (EUROSTAT, UNSD, UNECE, IATUR, …) as a valid and reliable source to study how people and households spend their time.

TUS measure the amount of time people spend doing various activities, such as paid work, household and family care, personal care, voluntary work, social life, travel, and leisure activities. Today, time use results are used to support the SDGs with comparable input.

Quest for TUS

Traditionally, household members are asked to write down their activities in a  paper diary. This process of data collection is seen as burdensome for both the respondent and the NSIs, turning into:

  • Low response rates
  • Large time investment from the respondent
  • High data collection and processing costs for the NSIs

At the same time globalisation is changing our world and a new digital transformation is taking place, turning into:

  • Quest to capture new activities, contexts and time use patterns
  • Quest to improve the quality of the data collection, and much cheaper and faster

Trusted Smart TUS

To support national and regional policymaking new metrics are needed, enhancing new collection strategies that are intelligent and more interactive in the way society is measured.

Of key importance for Smart surveys is the interplay between active and passive registration, with the respondent as the central position:

  • Smart survey: surveys that make use of smart devices like a smartphone, wearables, smartwatches, … and extract extra information from the behaviour of people
  • Active registration: straight-away input of the respondent to a front-office application (web/mobile) portal
  • Passive registration: inclusion of external sources of information (sensor, administrative databases, …) where the respondent does not have to do anything (but consent)
  • Interplay: external input that supports the respondent task to register qualitative answers but the other way around to provide the means to the respondent to adjust and supplement the input from external sources

The next step is to let TUS benefit from the technological developments in order to support the data quality and to lower the time con respondents.

Geolocation project objective

A basic strategy in TUS is to record activities within a temporal, spatial and social context. The idea behind is the improved registration quality when respondents remember the context within which an activity took place.

The strategy is based upon a modern version of the Hägerstrand’s time-space diagram (also called a framework) where household members spend a day of time interacting with the place and social environment.

The objective is to visualise the places where the respondent or household members have been throughout the day and which modes of transport they have used.

Creation of a plugin

Development of a Cordova Background Geolocation plugin for iOS and Android, capturing passive data:

  • longitude & latitude
  • time stamp
  • activity: still, walking, running, bicycle, vehicle
  • extra information on the location via connection with external database (e.g. Foursquare)

Plug and play principle

Transmission of the plugin data to a plugin-server. The plugin server can be called by candidate platforms resulting in efficiency and productivity gains through collaboration in sharing tools and infrastructure.

Integration into MOTUS

A visualisation of the space-time diagram is provided to the respondent via a web and mobile application in such a way that the passive data input can be accepted, changed and/or supplemented in an active manner.

The interaction between the passive and active input leads to Trusted Insightful Smart TUS.

SOURCE TM

Posted on August 31, 2018 by -

The motivation

In 2017 EUROSTAT started a project to modernise the collection of Social Statistics (ESS – European Social Statistics). By doing so, EUROSTAT builds on the Wiesbaden Memorandum from 2011 that strives for better information about patterns of time-use and consumption of households. More specifically, it involves the time-use survey (TUS) and household budget survey (HBS).

The project

Against this backdrop the SOURCE TM project emerged. SOURCE TM stands for Software OUtreach and Redefinition to Collect E-data Through MOTUS. The central aim of this project is to collect time-use data in Europe in a comparable way. Apart from the Research Group TOR, both STATBEL (the statistical office of Belgium) and DESTATIS (the statistical office of Germany) are involved. Within this project they will expand their knowledge about the MOTUS software platform, focusing on the way data is collected through a MOTUS application that runs both online and offline.

The challenges

The major challenge for the European modernisation project regarding the collection of Social Statistics is striving for comparability while at the same time leaving room for country specific interests (f.e. in terms of questions or activities) and wishes or concerns (f.e. in terms of the influx of respondents or the length of the survey). The MOTUS software platform accepts this challenge, because:

  • MOTUS is developed to design all survey components (questionnaires, diaries, context) as well as all communication with respondents within a single program (i.e. comparability); and
  • MOTUS is capable of adding unique contextual elements to this program at the same time (i.e. country-specific interests/wishes).

The quest for a proper configuration of MOTUS entails two phases of testing:

  • Developing a prototype of an e-TUS (online time-use survey) that will be evaluated by a large groep of scientists and representative that form the Workgroup TUS of EUROSTAT.
  • Having a non-representative sample registering their time-use for two days (one weekday and on weekend day).

“Comparability and customisation are two important conditions for the European modernisation project of collecting Social Statistics to be successful. Both elements are at the base of MOTUS.”

Both TUS and HSB are based on registrations in a diary: TUS for time-use and HBS for household consumption. Logically, both STATBEL and DESTATIS posed the question what it would take to use MOTUS for HBS purposes as well. Since the programming of MOTUS allows for such adaptations, this question has been included as an additional line of research in the project.

The aim

If successful, the aim of the project, which runs from January 2019 till February 2020, is to include MOTUS as a method for online time-use surveys (e-TUS) and online household budget surveys (e-HBS) in the CSPA catago. CSPA stand for Common Statistical Production Architecture. It describes that standards and principles for the production of national statistics and is aimed at improving the comparability of results.

BEHAVE: setup of a behavioural panel through MOTUS

Posted on August 31, 2018 by -

Background

Online panels are unavoidable in science

An individual’s opinion is central to better understanding the needs and behaviours of customers, users of services, and employees. The power of opinion is getting more important in market research. The internet provides a quick and cheap way to collect opinions, which gives online panels an important role in the collection of market oriented data (see regulations ISO26362). Fast and cheap available data also leads to the increased use of online panels for scientific purposes.

How reliable and valid are online panels?

Companies exploiting panels parade with the size of their panels and the subpopulations that are part of it. These two elements support the service they provide: opinions of a representative sample of the population or group in question.

On the contrary, information about the recruitment or entry of panel members is often scarce. This information, however, is essential for knowing the (non-)response rate and make generalised statements about a (sub)population. More information, for example about how often panel members participate successfully and about the quality of their answers, is often lacking as well.

A prerequisite for reliable and valid results is a random sample (f.e. from the National Register). This increases the representativeness of the panel and supports generalised statements about a (sub)populations. At the same time, all decisions and steps in the process of creating a panel need to be documented.

International examples of academic panels

There are a small number of panels in Europe that are managed according to academic standards. Examples are

  • LISS panel – Langlopende Internet Studies voor de Sociale wetenschappen – The Netherlands
  • GESIS panel – Leibniz-Institute für Sozialwissenschaften – Germany
  • ELIPSS panel – Étude Longitudinale par Internet Pour les Sciences Sociales – France

Outside Europe, some leading panels exists. Examples are:

  • HILDA panel – Household, Income and Labour Dynamics in Australia – Australia
  • ALP panel – American Life Panel – America
  • UAS panel – Understanding America Study – America

It is striking that most panels are oriented towards social sciences with a focus on the individual. The HILDA panel is an exception. This panel focusses on the household from an (socio-)economic point of view.

Project BEHAVE

Design and aim

Project BEHAVE combines the need for an academically oriented panel with the focus on expanding knowledge on longitudinal patterns of behaviour. Data on human behaviour will be collected using the MOTUS software platform, both by means of active and passive registration. The project includes an interdisciplinary team and has the following goals:

  • Creating a panel in Belgium according to academic principles of reliability and validity;
  • Linking this panel to MOTUS, which will serve as the respondent management and research coordination platform;
  • Expanding existing methods of active registration (i.e. active involvement of respondent) with methods of passive registration (i.e. using wearables, sensors, and databases linked to MOTUS);
  • Focussing on a longitudinal, multidisciplinary, and open source data collection strategy.

Together, these aims make BEHAVE a unique project.

The process

The BEHAVE project exists of three types of partners. Central to the project is a multidisciplinary team of scientists at the VUB, from the department of Social Sciences & Solvay Business Schools, the department of Engineering Sciences, the department of Medicine and Pharmacy, and the department of Physical Education and Physiotherapy. In addition, an open call is done to include other scientific institutes. Finally, third parties can participate as well, as there might be: policy institutions, ngo’s, non-profit organisation, companies, etc. At this moment, 23 third parties have expressed their interest, including the Belgian statistical office (STATBEL).

“The research community needs an academic panel to render reliable and valid opinions of people. This holds both for scientific as well as for market research.”

The academic principles of reliability and validity are priority when creating the BEHAVE panel, as is securing privacy of its panel members.

Aim: at least 5,000 to 10,000 panel members

The project aims to include 5,000 randomly selected respondents in a panel in the first phase of the project. Behavioural research requires a greater effort of respondents than opinion research. This means that large scale studies can at most be repeated every three months. This can be supplemented with studies of a smaller scale. In a second phase, the project aims to increase the size of the panel to 10,000 respondents. Based on the interest and inclusion of other institutions, this number might rise.

Learn more?

The project starts in October 2018 and runs for four years. If you want to learn more about the project and the possibility to be included, contact Joeri Minnen.