Financing: RRF-2.3.1-21-2022-00004 Artificial Intelligence National Laboratory (MILAB), Proof of Concept (HUN-REN)

Project start: 2021

Project duration: continous

 

In recent years, face-to-face scientific data collection has been in crisis worldwide, including in Hungary. Every day, empirical social research is confronted with the fact that this method is becoming increasingly difficult and costly to implement. As a result, non-face-to-face data collection is becoming more and more prominent, and at the same time research that (also) uses 'digital footprints' of human behaviour, which can replace or complement certain questionnaire-based surveys. This data can come from a variety of sources: Website traffic, social media activity, app usage and data on movement and other physical conditions from various sensors and log files on devices. These are non-intrusive methods of data collection that do not require constant user interaction, but provide a high level of accuracy and detail about people's digital lifestyles. Combining different passive data collection techniques with the survey method, i.e. making the activation of different questionnaires conditional on the fulfilment of certain events (e.g. displaying a questionnaire block when a user is in a certain location or when a certain amount of screen time or app usage is reached), can open up new perspectives.

 

Octopus Research Tools (ORT)

 

As part of the National Laboratory for Artificial Intelligence (MILAB) and with the support of the HUN-REN POC grant, the development of Octopus Research Tools began in autumn 2021 at the Social Science Research Centre with the participation of social scientists and computer scientists. The aim of the Octopus project is to combine these two research methods, i.e. sensor and device usage protocol-based digital behavioural data collection and surveys that can display multimedia content (audio, images, video), in a complex smartphone-based software ecosystem. The aim was to create an innovative research tool for scientific and even commercial investigation of a wide range of social issues and problems, for modelling and predicting human behaviour and, on this basis, for supporting policy decisions.

In its current form, the Octopus system consists of three large units:

 

1) Users connect to the system via an application that allows them to track the tasks they have been given in each research project (e.g. completing a questionnaire or uploading a photo) and to see in a transparent way the amount of data collected by the application, the level of detail of the data and the process of anonymisation.

 

2) The Octopus central server is directly connected to the HUN-REN network and stores the data securely and reliably. An application installed on the mobile phones sends the data recorded on the phones to the Octopus central server at regular intervals via an internet connection and checks whether there is a new data collection task in the survey. One of the advantages of Octopus over other currently available data collection solutions, especially questionnaire software, is that no permanent internet connection is required for data collection.

 

3) In the research administration interface, researchers can plan their research, develop questionnaires or "fine-tune" the way data is collected for each module, for example by setting the frequency of location recording or the accuracy of the location data collected.

 

The general framework of the software and basic rights management was finalised by mid-2023. Passive data collection modules include a complex project editor interface, a questionnaire editor (with single-choice, multiple-choice, text and slider questions), trigger and notification modules for the time of data collection, sampling, location and application usage. One of the biggest advantages of the application is that it can be used both offlineand online, so that no constant online connection is required to complete the questionnaires.

 

The smartphone app is currently available for Android devices via a direct (APK-based) installation. The iOS version works via the Testflight app. It is planned that both software will be available in the public stores, but approval by Google and Apple will be subject to a longer approval process due to the sensitive data collection solutions.

The data will be stored on the Centre for Social Science Research's own servers from mid-2023, when the first test studies are carried out. However, the system can of course also be used on the servers of external partners, as this is the way to ensure its widespread use and subsequent commercialisation.

 

More information: www.octopus-research.hu