Mapping Haggadah printed editions (and being careful about it)

The bibliography of the Hebrew Book (BHB) database is an essential reference tool for Jewish studies, made available by the National Library of Israel. Yet, similar to many other library resources,  it functions mainly within the limited search paradigm, in which the user is always conceived as looking to find a specific book,  or several books, and the user interface is there to help in the retrieval. But the valuable data collected, years of expert work that were put into this knowledge base can also be used differently when liberated from the search interface and made available for enrichment, analysis and visualization.

In the framework of the project DiJeSt, with Dr. Yael Netzer and Dr. Kepa J. Rodriguez,  we prepared the Bibliography of the Hebrew Book database for “Distant reading” by processing date and place of publication, and modelling the data according to standard ontologies. For now, I am sharing a part of it as a passover treat, along with some notes on the promises and perils of distant reading datasets.

In the following map you can see all the places documented in the BHB where Haggadah editions were printed, with their languages of translation. By pressing the side button, a legend will open that will help you explore. Clicking a pin will also open the details of an edition published in that place.

When data is available for distant reading, we can often see historical, geographic and other patterns that are more difficult to perceive when looking at records more closely.  Interesting outliers are also more visible this way: why were English translations printed in Hannover, or Fuerth? or a Judeo-Arabic translation in Vienna? 

Each of this colored points that catches the eye may be a trigger to exploration of fascinating stories. Some of these stories, dealing with the translation to English, make the subject of Avraham Roos’ dissertation, and along with other experiments with knowledge visualization of the Haggadah translation phenomenon, can be read on his website.

Visualization is never a ‘view from nowhere’: it always conceals as much as it unveils and it is important to remain critical towards it, or complement the view it provides with alternative angles. This is an important caveat: a visualization of a database is always no more than that, and usually less;  first,  as thorough as the BHB may be, it does not cover all Haggadah editions ever printed. Whether it is because Haggadah editions simply didn’t leave any trace to be recorded, or because of the documenting policy of the BHB project. On top of this, the retrieval method I used dropped several Haggadot: I only took those that have a Yudilov’s Otzar HaHagadot identifier, and have an exact (settlement) level location that can be mapped. Dozens of Haggadot are only located in estimated regions, such as “Poland/Russia”, and they will not appear on the map. Finally, one should always remember that one edition that was documented in the BHB could have been printed in changing numbers of copies; records of printing therefore do not represent records of reading and usage. To add to the dataset criticism, one should be wary of taking the visualization as a straightforward representation of the data:  In both maps above, several editions published in the same place would only be represented with one pin. One edition – or Hundreds of Haggada editions – may have been printed in many of the places marked here with one pin.
This last limitation, however, can be mended by an alternative visualization: to express  the number of editions over time I am using the tool Palladio,  to which one can upload geo-temporal data and explore it through a map, a timeline, and several other facets and visualization. The passage of times and the numbers tell their own story. This is what a travel through time in the BHB Haggadah record collection looks like:

What does this view reveal? and just as important a question: what does it conceal?