The goal of these three tools is the same to search and assemble large amounts of data into formats which users can easily manipulate to create visualizations of their data. The goal of these tools is to enhance research and help researchers provide hard data and evidence for their humanities projects. The tools help us shape new questions and provide additional options for the portrayal of “ideas of change”, show how people, ideas, organizations or events have common patterns and interactions all of which enhance the meaning and understanding of the data and the humanist’s ability to teach and share the insights from it.
Voyant focuses on text and identified collections. It inspires thinking into the nuanced meaning and use of words. This includes the “power words” within a text and also the auxiliary words which demonstrate a deeper level of language use in a particular context. Although targeted for the “new to digital technology humanist” and straightforward in its process; if one is really new to the processes and navigates off the script or attempts to add something, there are no ‘aids’ built into the system to help you easily answer questions or solve problems.
CartoDB is a network mapping tool or as Carto describes itself, “An open, powerful, and intuitive platform for discovering and predicting the key insights underlying the location data in our world.” (https://twitter.com/cartodb?lang=en) This is a more sophisticated program for a new user than Voyant because it has more “built in software” making it more accessible. CartoDB focus on the spatial aspects of the data – how it is related geographically. Layers can be added with new information to enhance the understanding of the connections between relationships. It is easy to switch between and add maps. Starting with basic geography, but using GIS reference and point placement, CartoDB makes the maps for you. It also projects the maps in different ways using widgets: ‘heat maps’ show concentrations of activities and ‘animated’ maps can demonstrate interactions over time and space, or demonstrate patterns and isolations. This map is very helpful for clear comparisons and relations.
Palladio describes itself a s a “platform” for network analysis. This platform has a number of tools and is organized around manipulating the corpus data. When I was using this platform, I had the feeling; I was using a very limited aspect of its capabilities. “It is designed to manipulate data the way historians think” this is useful for us historians who have trouble “organizing our research” in the same ways that Scientists and Social Scientists do. Palladio enables the contents, of seemingly incomparable elements, get turned into plotable graphs, maps and clouds (spatial comparisons for ideas and categories). Being able to manipulate the data without having to have the underlying programming expertise is magic. However creating the corpus of data becomes highly significant, more on this below. The user has to ‘trust’ the logic and ability of the programmer (that the description of what is being done with the data – is being done in the same way the historian “thinks” it is being done). Decisions are made regarding how to organize the data sets. Errors can result in huge consequences for interpretation.
When computers were first coming into popular use (1970’s & 80’s) the key phrase was GIGO (Garbage In/Garbage Out). Manipulating research is pretty powerful and the visualizations ‘lock in’ ideas more definitively than ‘discussion’ does. As I read about and worked with these tools, I was excited about how a project I am working on Women and Health Care in Cleveland would benefit by their use. The “tools” will provide quick ways to visualize the impact of location and intersecting relationships we have been struggling to describe. Each tool can help us see the data from a new perspective and solidify, highlight our findings or indicate new questions. But, these tools cannot fully explain all the “whys” of the locations; the personalities, beliefs, or causes of actions. Missing records from whole segments of Cleveland’s population would not be “factored in”. As I worked with the individual tools, it became evident how each of the projects that used them, needed to modify them to suit their own data and questions. The itemization of the “Grant Support” is evidence of the need for modification, and support for programmers and editors to work with the researchers so that the questions are being addressed thoroughly and potential problems recognized and documented so that we do not generate (in today’s vernacular) “Fake History”.
Utilizing multiple tools is one way to avoid egregious errors and explore the data more deeply. Use of Voyant on the text of the Slave Narratives reflected the use of language which evidenced the fact that the slaves being interviewed were mostly plantation based and isolated. CartoDB demonstrated that “Alabama” interviews were not conducted all over Alabama, but in pocketed areas. It also provided evidence that the numbers of interviews peaked in 1937 (but the data did not indicate why). The similarity and limitations of the script used in the interviews was evidenced in Palladio by a cloud diagram of topics. These tools provide amazing insights, enable analysis that would not be possible without the speeds of text manipulation and they are great aids to the craft of history.
