Community Research and Development Information Service - CORDIS


CIMPLEX Report Summary

Project ID: 641191
Funded under: H2020-EU.1.2.2.

Periodic Reporting for period 1 - CIMPLEX (Bringing CItizens, Models and Data together in Participatory, Interactive SociaL EXploratories)

Reporting period: 2015-01-01 to 2015-12-31

Summary of the context and overall objectives of the project

The project aims to develop modeling, computational, and ICT tools needed to predict and influence disease spread and other contagion phenomena in complex social systems. To achieve non-incremental advances we will combine large scale, realistic, data-driven models with participatory data-collection and advanced methods for Big Data analysis. In particular we will go beyond the one-dimensional focus of current approaches tackling one aspect of the problem at a time. We will interconnect contagion progression (e.g. epidemics) with social adaptation, the economic impact and other systemic aspects that will finally allow a complete analysis of the inherent systemic risk. We will develop models dealing with multiple time and length scales simultaneously, leading to the definition of new, layered computational approaches. Towards policy impact and social response we will work to close the loop between models, data, behavior and perception and develop new concepts for the explanation, visualization and interaction with data and models both on individual and on collective level. We will cast the fundamental advances into an integrated system building on widely accepted open ICT technologies that will be used and useful beyond the project. As a tangible ICT outcome directed at facilitating the uptake and impact of the project, we will implement “Interactive Social Exploratories” defined as interactive environments which act as a front-end to a set of parameterizable and adjustable models, data analysis techniques, visualization methods and data collection frameworks. In summary, we aim to (1) produce fundamental theoretical, methodological and technological advances (2) mold them into a broadly usable ICT platform that will be a catalyst for producing, delivering, and embedding scientific evidence into the policy and societal processes and (3) evaluate the system empirically with policy makers and citizens focusing on the concrete problem of epidemic spreading.

Work performed from the beginning of the project to the end of the period covered by the report and main results achieved so far

Work in PY1 has focused on (1) furthering the theoretical understanding of complex systems necessary to predict and influence general spreading phenomena in social-technical systems, (2) determining appropriate conceptual and computational approaches to suit this, (3) developing basic methods for privacy aware social mining and distributed, adaptive data collection, (4) implementing an initial version of a visualization framework for the explanation of scientific evidence, (5) initial concept oft facilitating mass participation in data collection and influencing social contagion, (6) progress towards multiscale modeling platform for disease spread and (7) the development of the overall system architecture.

Progress beyond the state of the art and expected potential impact (including the socio-economic impact and the wider societal implications of the project so far)

Our analytical framework represents the complex socio-technical systems we model as multiplex networks with nodes characterized by relevant state variables, and links that represent direct interactions between nodes. The networks are multiplex in the sense that nodes have different types of connections (or layers), such as those representing face-to- face interactions or those representing online communications. Importantly, these links are not static, but may vary in time, for example when representing social interactions or the movements of people. We will highlight the importance of the dynamical aspect of networks throughout the theoretical results we present below. The temporal and spatial resolution of these networks is flexible and depends on the specific situation to be modelled. Furthermore, multiple resolution scales can be weaved together to model a specific situation. Spreading processes are a change in state variables that can be understood through the interaction structures. In this representation, the multiscale geographical approaches represent a key part of the research exploration. With a framework that can factor in social/behavioral feedback at different spatial and temporal resolutions we expect to more accurately predict how a spreading process unfolds, but also to simulate the consequences of different informational interventions.

To provide data to calibrate and validate the models a novel concept of social interactions based, distributed data collection and aggregation has been developed. The concept is based of the notion propagating request for data along people’s social networks with everyone deciding, which request to propagate based on his/hers individual assessment of the importance of the request. In a similar manner collected data is incrementally aggregated, interpreted and propagated back to the source of the request in a way that respects individual policies and privacy concerns. The concepts is currently being implemented in mobile App that is connected to the Influenza Net (which is a production data collection system for monitoring influenza spread).

In parallel the development of new social mining and methods connecting the data to the models and capable of adapting to changes in model representation and configurations has started This includes on one hand methods based on multi-dimensional network representation applied to the specific problem evolution of social communities and interplay between individual profiles and collective patterns. On the other hand, data mining process that addresses a particular formulation of the link prediction problem for dynamic networks, called Interaction Prediction has been proposed. New data based models of human mobility taking into account so called returners and explored have also been studied.

With respect to the explanation of scientific evidence an initial prototype has been developed that consists of multiple linked views that allow for a visual explanation and verification if mobility patterns extracted from social media align with explicit movement data (e.g. public transportation, flight schedules, etc.). The analysis can compare different datasets on a raw data level, as well as on an abstract level (graph model with hierarchical aggregation).

Towards a modeling platform for interactive multidisciplinary modeling a data mining query language, PL/DMQL capable of manipulating data, to extract model and patterns, store and combine them in order to build more complex analytical processes has been defined.

To facilitate new concepts for mass participation in data collection, analysis and decision making processes workshops with the partners from ETHZ and DFKI highlighted key aspects for motivating individuals to contribute to data collections are personal benefits and the protection of individual-related data (T4.2). Currently we are discussing mass health data collection and focus on the possible realizations in the form of web service

Related information

Record Number: 186463 / Last updated on: 2016-07-12