Measuring Social Behavior via Dynamic Network Interaction

This project will (Aim 1) build, test, and implement a cost-effective mobile/wearable cross-platform continuoustime interaction (CTI) data acquisition system operated via (Aim 2) a web-based administration service. Thiscombination of a mobile app and administration service software will allow future social and behavioral healthresearchers to administer their study?s behavioral and social tracking protocols and manage the collected CTIdata remotely, across significantly larger social research scales, and to do so either locally or in the cloud. Thenew set of tools will combine ultra-fine-grained social network tracking with cutting-edge instantaneous timesampling to allow for ?thicker? description of captured interactions. To complement the app and admin service,the project will also implement (Aim 3) the first extensible suite of ?continuous interaction? data analysis tools inthe open-source cross-platform analytical environment R, which public health researchers can use to analyzethe CTI data collected in the course of their studies. The final product, including all of these elements, willconstitute an Open Dynamic Interaction Network (ODIN) platform, which will be widely available for public usewithin 3 years of the start of the project. To validate the outcomes of the development and implementation process of Aims 1 & 2 for a broad rangeof possible uses, the project will test the ODIN App and web-based ODIN Admin Service via test-bedimplementation using a mobile technology lab at the University of Nebraska-Lincoln. These tests will allow theresearch team to validate the integration of software and hardware systems, test their robustness, usability,and flexibility, towards finalizing the deployment specifications of the platform as a whole. To test the ODINAnalysis suite, synthetic CTI data generated by the MABUSE simulation platform (previously developed by thePIs (via RC1 DA028476-01) will be fed into the analysis software in order to validate its capabilities, andquantify and optimize its robustness to noisy input. Taken together the above battery of tests ensures that wecan validate ODIN?s CTI data collection capabilities and the ability of the ODIN analytic software to reliablydiscover and model diffusion and selection processes in social and bio-ecological data. As part of ourcommitment to see this technology used widely in the understanding and promotion of public health, in the finalyear, the system will be made deployable in the cloud (over Amazon Web Services, Google Play, CRAN, andGitHub), under open source licenses. Increasing the accuracy and fidelity of dynamic network data collection is of growing importance as basichealth science research looks to human interaction and social network analysis to explain health dynamics. Ofthe currently supported R01-level program announcements at the National Institutes of Health, 17 explicitlyidentify ?social networks? as important foci of research. As such, ODIN represents a significant contribution tohealth science more generally, with a potential impact across a wide range of project implementations.