Abstract:
|
As data are readily available at our fingertips, it has become common practice to integrate, blend, or link data. This is a classical problem that derives from survey methodology known as record linkage or entity resolution, which merges together noisy databases and removes duplicate entities. These methods were created in survey methodology to identify the size (for example, members of a household). In return, resources, such as schools or roads, could be distributed to communities. In this roundtable, we will discuss how record linkage has changed with the rise of big data, computational advancements, and tracking devices. Is there still a need to integrate data sources for the public good? Are there situations where such methods are not in the best interest for society? Specifically, are there guidelines that should/could be put in place to mitigate such emerging issues and setting best practices? How could we improve the current diagnostics and visualizations toolboxes for data integration?
|