Increasing access to the web of "broad data"

Abstract

Due to a problem with my wrists, hopefully temporary, I am writing this short article using voice recognition software. As this is the first time I've used this software, I'm actually relatively pleased with how well it works. I am able to write e-mails, send twitter and Facebook updates, and do many routine tasks using my voice. Modulo a few typos, I can even write this document (although putting this into ACM format required help). However, a significant portion of my life nowadays involves interacting with data, and when it comes to data interaction, "web for all" may as well just be a slogan. Even without accessibility problems, whether your goal is to enter, discover, or integrate data, or to try to understand what some particular data is telling you, it isn't easy. Add in disability, and the problem is made much worse. In this keynote, I will discuss some of the issues that arise as people try to use the "broad data" that can be found on the World Wide Web. The modern combination of "lightweight" semantics, based to a large degree on the rapidly maturing products of early semantic Web research, coupled with the "big data" tools that have moved away from traditional relational databases, provides an area of exploration that is pushing research in new and interesting directions. Tim Berners--Lee's call for "Raw Data Now" is being heeded in many quarters, and other forces, including those of transparency and innovation, are creating vast repositories of data that are available without restriction. As an example, governments around the world have been posting data sets on the web at a really amazing rate. In the past year and a half, my research group has identified and indexed the metadata for well over 700,000 open government datasets from around the world. This includes, at the time of this writing, datasets from more than thirty countries and international organizations in 16 different languages (we currently anticipate having more than 1 million data sets by the end of this calendar year). Our research has explored how to create, index and search metadata from this immense Federated catalog space. We have also been developing tools for helping users to create linked data from these data sets and to use that link data in the development of visualizations and other presentations that make the data more accessible. We are also working with the US government on bringing these techniques to the US Data.gov project. In this talk, which I admit includes parts that are far more speculative than practical at this point in time, I will explore how the link spaces among the data provide the underpinnings of potential new applications that will help bring data analytics into our personal lives. By making data more personalized, we may be able to achieve new possibilities in data integration that could provide capabilities by which all of us would be able to more fully interact with the important data that affects us in our everyday lives (such as health and well-being), and not just in our professional careers. I contend that similar techniques could be used to help increase the accessibility of data on the web. Linked-data approaches have been helping to some degree in this arena, but still leave a lot to be desired. In short, I will explore some exciting things happening on the web of data, but bemoan the challenges that still remain in providing scalable access to the Web of Data.