Merging statistics and geospatial information, 2016 projects - Greece

Integration of geospatial information and statistics. Development of an application to make geolocation enabled e-questionnaires available to mobile devices of census enumerators; 2016 project; final report October 2018

This article forms part of Eurostat’s statistical report on the Integration of statistical and geospatial information.

Full article

Problem

There was not a complete and fully operational cadastre and therefore no complete and coherent georeferenced buildings register. A more accurate and timely recording of the population including their dwellings’ location and other reference data was needed, for example for the 2021 census.

Objectives

The aim was to develop an application that could make geolocation enabled e-questionnaires available for use on the mobile devices of census enumerators. The solution should:

have the functionalities to design and distribute e-questionnaires including enhanced geolocation capabilities,
have real time validations/checks,
link address registers to actual locations,
store data locally (in the mobile device) and send data to the server,
serve as a prototype for designing future mobile questionnaires (for example, for the 2021 census),
be as independent as possible of mobile software variations,
follow open standards,
be hosted on the infrastructure of the Greek statistical office.

Method

The software application platform was developed using open-source programming languages and frameworks. It was composed of a few parts.

A designer module. This was developed to:
- set up a new survey and e-questionnaire;
- create and assign user roles as well as create user accounts;
- create entities within the e-questionnaire, declare the fields for each entity and define the relationships between entities;
- develop functionalities such as data importing, validation and the flow of the questionnaire;
- design the look and feel of the questionnaire;
- approve the survey as being available and then as being published.
A portal. This was developed to give supervisors the functionality to review and control the progress of a survey. For example, supervisors can select a survey, identify and evaluate the performance of enumerators, replace enumerators, assign census blocks to enumerators.
Android application. This is created when the design of the e-questionnaire is approved; the e-questionnaire is then published. The application:
- can be downloaded by enumerators and installed on a mobile device;
- is used to
  - complete the survey (for a respondent)
  - temporarily store the results on the mobile device
  - transmit the results to a server of the statistical office (and delete the temporary storage from the mobile device).

The Android application was tested on three smartphones and one tablet. Some were new and some used, had different CPUs, chipsets, storage, screen sizes and operating systems.

The survey used in the study was composed of a questionnaire collecting information about a building and a second one collecting information about individual households within the building. The questionnaires and validation checks were based on the 2011 questionnaire.

The test was conducted in four areas of Piraeus (city centre and outskirts) as well as in two villages – Katalakkos and Propoulion.

During the study, the coordinates of the main entrance of a building were used in the e-questionnaire as this location was also used in the Hellenic Post address file. Enumerators could use a drop-down list to select an address from the Hellenic Post address file or add a new address if needed.

Results

From a functional perspective, the collection of data using smartphones was successful with no issues related to battery life. Use of the e-questionnaire was more troublesome on smartphones with screen sizes of 4.5 to 4.7 inches or smaller. It was easier to use on the tablet (with a 10-inch screen).

The survey in urban areas resulted in a limited number of corrections to the Hellenic Post address file. High buildings impacted on the accuracy of GPS signals. The survey in rural areas faced other issues, such as not always being able to approach a building’s entrance if a property was inaccessible, for example because of a locked gate/fence. The use of the Hellenic Post address file was successful in areas for which it has data. In other areas, typically rural ones, the use of e-questionnaires on a smartphone proved to be a successful way to collect geolocated building and household information.

Figure 1: Katalakkos settlement potential buildings/locations along with the collected locations from the tablet questionnaire
Note: the highest precision in location is in the open areas denoted by two red circles (school with schoolyard in the upper right part, church building in lower left part).

It was concluded that data collection with mobile devices can be done at lower costs compared with traditional means, increases productivity and improves workforce management. Issues to be considered are the technology to be used, the need for training, the cost (for example of adequate mobile devices), data security and the speed of technological change.

The study provided an opportunity to gain practical experience in setting up such a survey from the perspective of a survey designer, a person responsible for approving questionnaires, enumerators and supervisors.