Home

Introducing IDEANet - Integrating Data Exchange and Analysis for Networks

Over the last 40 years, network analysis has emerged as a prominent approach to data-intensive research. Despite this steady growth and investment, network analysis remains something of a niche specialty that can exclude novice users who usually only receive standard statistical training. Moreover, since much of the growth in network science has tended to be field-specific, tools and formats have developed independently across disciplines. The multiplicity of formats and sensitivity of social data makes existing records difficult to share across scholars in the field, limiting the opportunity for new findings on the already accumulated body of network data. IDEANet - Integrating Data Exchange and Analysis for Networks - aims to maximize scientific discovery in human network science by significantly lowering the analytic and access barriers-to-entry for researchers. IDEANet is supported by the National Science Foundation as part of the Human Networks and Data Science - Infrastructure program (BCS-2024271 and BCS-2140024).

IDEAnet features three key components (1) a suite of analysis tools developed in R which automatically generate standardized network analytic measures (2) a GUI (Graphical User Interface) which gives access to the aforementioned measurements through an easy-to-use menu-based program and (3) a secure data repository that routinizes the capacity for archiving and accessing network data, including sensitive data.

The analysis tools are distributed as a package and built with real-world data constraints in mind to allow novice users the ability to gain substantive results as efficiently (but still accurately) as possible. Core metrics comprise 17 node-level measurement (e.g., degree, centralities, reachability) in addition to 27 system-level metrics (e.g., network size, dyad census, transitivity). Additional modules include multiple regression QAP, multi-relational blockmodeling and a community-detection routine that partitions the network based on 10 commonly used methods and evaluates their concordance using CHAMP. Further modules are in development including meta-population disease simulation and dynamic network diffusion simulation.

The secure data repository is hosted on Dataverse in collaboration with Duke University Library. Researchers are often interested in sharing their data but can be limited by strict institutional requirements. Our repository facilitates this transition by offering three levels of data security: Open access, Secure non-restricted, and Secure use-restricted. Secure non-restricted data include some level of confidentiality such that investigators require IRB approval for access. Secure use-restricted data requires both IRB approval for use and further substantive limitations required by the data owner. To accommodate the diversity in requirements, IDEAnet makes use of the imPACT architecture – a “notary service” that seamlessly matches user certification and data access requirements.

The difficult learning curve involved in learning network tools means that researchers with substantive interests in network processes but who are not specially trained in network methods must either invest heavily in training or risk serious analytic errors. The goal of IDEANet is to provide an integrated network data analysis framework within R that capitalizes on the best of current tools while building robust safeguards against common data and analytic errors.

This website serves as the home for IDEANet and contains vignettes describing how to use the package in R, navigate the GUI and access the data repository. If you have any questions, feel free to use the contact information located in our Contact Us page.