Tag Archives: API

Visual Tools: Voyant, CartoDB, and Palladio

New web-based, open source technology has dramatically shifted the landscape of digital humanities. It has affected fields related to  digital humanities in two significant ways. For institutions and digital humanists a new quest to create, build, and host project sites has emerged. These digital projects allow users to interact and manipulate data in specific ways that yield almost infinite combinations. For users, these digital projects have laid the groundwork for moving research beyond the archive and to digest and draw conclusions based on datasets and information expressed through new macro-based visuals. The projects/programs reviewed here focus on textual analysis, geospatial mapping, and visual graphing based on large sets of metadata and archival information.

Strength/Weakness: The strength of Voyant is the range of text analysis provided: cirrus, networks, graphs, context, verbal patterns. This is also its weakness. At first glance it’s very impressive but when trying to set or manipulate certain features available to the user for the purposes of customization or multiple datasets, the program does not function well.
Similarity/Difference: Voyant is similar to CartoDB and Palladio in that they are all free open-source, web-based programs. Voyant and Palladio do not require usernames or passwords. Voyant is different from CartoDB because CartoDB does require a sign-up. Voyant is different from Palladio because Voyant has one main screen with several visual fields, while Palladio only focuses on one type of visual analysis at a time, i.e. maps or graphs.
Complement: Voyant provides sophisticated text analysis and CartoDB provides sophisticated geographical analysis. Paired together, they provide unbelievably rich yet simple ways to “see” data relationships. Palladio and Voyant complement one another because they allow users to layer and filter the same data to produce different types of word graphs, clouds, and networks.

Strength/Weakness: The strength of CartoDB is the visual clarity and graphic options for its maps. The program’s weakness is that it really only serves to create maps and not graphs or other visual organizers. As a side note, this could just as easily be a strength because it does one thing well.
Similarity/Difference: CartoDB is similar to Palladio in that it focuses on one type of visualization, which it does very well. It is different in that its foci are  maps=CartoDB and graphs=Palladio. CartoDB is similar to Voyant on a basic level; they both produce visual graphic representations of the relationships within a large set of data. They are different because Voyant attempts to do many things (but not geospatial mapping), while CartoDB focuses on geography and not text.
Complement: CartoDB and Voyant complement each other well for the same reasons that they differ (above). Voyant does what CartoDB does not and vice versa, so together they provide an even more comprehensive picture of patterns that can be draw from data. Palladio and CartoDB complement one another because each does a different thing well. I would be tempted to use these two rather than Voyant because they are both user friendly.

Strength/Weakness: The strength of Palladio is its relatively easy interface and the ability to drag and organize nodes and lines. The weakness of Palladio is the inability to save projects in formats other than svg or json, and that beyond the visual graphing network there is no additional information.
Similarity/Difference: It is similar to CartoDB in that it does have a map function, but Palladio is different because the most effective feature is visual network graphs. Palladio is similar to Voyant in that they both have word links and network features. They are different because Voyant is difficult to use (because of glitches not design), while Palladio is much easier to use.
Complement: Palladio complements Voyant by providing more options for word clouds and visual networks. Palladio provides a complement for CartoDB as they are both based on layering datasets manually with selecting different modes and filters.

As these open-source programs continued to “hone their skills” and “work out the kinks,”  they will no doubt provide continued and enhanced methods of data analysis that can be customized for and by individual interests.