• Introduction
• Presentation (ppt)
• Data & Methods
•  
• Query Tools
• Market Basket Analysis
• (a)Typical Readers
•  
• Texts
• Reading Level
• Topic Modeling
•  
• Conclusions

Introduction

In early December, 2011, we (the Humanities Digital Workshop at Washington University in St. Louis) first started using the data available from the What Middletown Read website. We turned to What Middletown Read in support of a research project undertaken by Lynne Tatlock, Hortense and Tobias Lewin Distinguished Professor in the Humanities at Washington University. Professor Tatlock's project is directed at understanding the extent to which German novels in translation were read in the United States in the late 19th century. Thus far, the research has yielded Professor Tatlock's recent monograph, German Writing/ American Reading: The Import of Women’s Fiction (1866-1917), and her presentation at "Distant Readings / Descriptive Turns: Topologies of German Culture in the Long Nineteenth Century," a conference recently held at Washington University.

Given our interest in late 19th century reading in the United States, What Middletown Read has proven invaluable. The website provides detailed circulation date from 1891 to 1902 (with a small, unimportant gap) from the Muncie Public Library, and is the result of an extensive cooperative effort between the Muncie Public Library, the Center for Middletown Studies at Ball State University and the Ball State Libraries.

The What Middletown Read website makes possible a thorough exploration of the data without any special computing or analytical facilities. Using just the website's features, we were able to quickly establish how often readers checked out the books we were interested in, what the characteristics of those readers were, and what other books they had checked out. In other words, just by using the website, we were able to find additional support for Professor Tatlock's hypotheses about the readership of German novels in translation.

The website's query facilities make it possible to search for information about authors, books, and library patrons. In cases where a query term (patron number, for example) allows a search across a range of values, the website allows users to in effect query for everything. For example, it's possible to query for all patron numbers between 1 and 999,999, and to have everything (or a very close approximation of everything) returned in the query results. Once the user has received a response to a query, the website permits him or her to download the results in a format convenient format for loading into a database, examining via custom processes, etc.

There is only one gap in the data made available from the What Middletown Read website. Using the website's tools for viewing query results, it's possible to view census information (age, gender, occupation) for library patrons. However, this data is not available for download. When we contacted James J. Connolly, Director, Center for Middletown Studies and Professor of History at Ball State University, about the gap, he generously made the entire database available to us.

In other words (and this is the point), What Middletown Read represents a remarkable example of open data. We would have not been able to do this work without the project's previous work and their ongoing generosity.