Giving shape to large digital libraries through exploratory data analysis
研究了探索性数据分析和可视化工具如何帮助理解大型书目数据集，并介绍了HathiTrust+Bookworm工具，该工具支持对HathiTrust数字图书馆中数百万部作品进行多维度探索。
Abstract The emergence of large multi‐institutional digital libraries has opened the door to aggregate‐level examinations of the published word. Such large‐scale analysis offers a new way to pursue traditional problems in the humanities and social sciences, using digital methods to ask routine questions of large corpora. However, inquiry into multiple centuries of books is constrained by the burdens of scale, where statistical inference is technically complex and limited by hurdles to access and flexibility. This work examines the role that exploratory data analysis and visualization tools may play in understanding large bibliographic datasets. We present one such tool, HathiTrust+Bookworm, which allows multifaceted exploration of the multimillion work HathiTrust Digital Library, and center it in the broader space of scholarly tools for exploratory data analysis.