Visualizing the Letters

During the annotation process that brought the Knowledge site to life, many data elements have been used to formulate descriptive paratextual statements about the Letters.

There are many alternative ways to explore semantic-rich data, each with varying levels of effectiveness. A flat index, albeit useful as a starting point to display and navigate data more efficiently, tends to collapse the depth of the data itself and to hide its most important characteristics. Instead, it may be more effective to adopt an approach based on multiple visualization strategies which support a much better sense-making of data through the use of charts and other visual elements that provide an accessible way to see and understand correlations and patterns in data.

What: the lexical terms

The pie chart and the donut chart are charts that convey a general idea of data that sum up to a total. They are efficient for displaying the relative frequencies of a small number of categories. Each category is represented by a slice and the area of each slice is proportional to the percentage of data instances belonging to that category. Pie and donut charts should have only a few slices (between 2 and 5), labeled accordingly with either values, percentages or both.

In the examples below, the donut chart on the left illustrates the frequencies of the five most frequently used classes in the dataset. It is immediately clear that doco:TextChunk, a class that belongs to the Document Components Ontology (DoCO) and that represents a piece of text defined by a start point and an end point, is the most frequently used class in the dataset.

The pie chart on the right shows the five most frequent types of doco:TextChunk that are present in the dataset. Attested name, which indicates a concrete occurrence of a person's name in a certain form, is the type of doco:TextChunk that has been used the most. It is followed by Philological note (a diplomatic-interpretative comment), Integration (a conjectural addition made by the editor), Lexical form (the technical lexicon used in the correspondence) and Change (a note that alters or substitutes a word or an expression in the text). In the dataset, each type of doco:TextChunk is related to its respective piece of text through the property dcterms:type.

The bar chart is a type of chart that displays data either vertically or horizontally by using columns to represent a category or attribute of that data. The height of each column reflects a frequency or a value for that category. Bar charts are useful for nominal comparison between many items belonging to a reduced number of categories. The horizontal display is useful when many categories are considered, since there is more room to accomodate the category labels.

In the dataset, each lexical form is an instance of doco:TextChunk with the property dcterms:type relating it to http://vespasianodabisticciletters.unibo.it/lexical-form. An instance of doco:TextChunk is related to its respective letter through the property dcterms:source, so it is possible to count the number of lexical forms per letter.

In the example below, the bar chart illustrates the frequency of lexical forms for each letter that contains at least one of them. The horizontal axis shows the number of lexical forms that are present in a letter. The vertical axis shows the letters, each identified by its respective label.

Where: the places of dispatch

A map chart is a chart that compares values and shows categories across geographical entities. It is used when specific geographical information is present in the data that is to be visualized, like coordinates, countries, regions, cities and so on.

In the dataset, the sending of a letter is modelled as an instance of pro:RoleInTime, a class of the Publishing Roles Ontology (PRO) which defines a situation describing a role a given agent may have in a particular period of time. Among other properties, the sending of a letter is characterized by being related to the role sender (identified by the URI http://vespasianodabisticciletters.unibo.it/sender) through the property pro:withRole and to a place through the property bio:place. Each place is an instance of bio:Place, a class that belongs to the Biographic Information vocabulary (BIO). In addition, the sending of a letter is related to the respective original exemplar of that letter through the property pro:relatesTo.

In the example below, the map illustrates most of the places from which the letters have been sent. Each icon marks a specific place, identified by its latitute and longitude. By clicking on an icon, a tooltip appears on the top of it. The tooltip shows the number of letters sent from that place as well as the full list of those letters, each identified by its label. Note that some places have not been considered in the visualization: namely, an unknown place (identified in the dataset by the URI http://vespasianodabisticciletters.unibo.it/unknown-place), from which the Letter 43 has been sent, and the Aragonese campsite (identified in the dataset by the URI http://vespasianodabisticciletters.unibo.it/accampamento-aragonese-location), from which the Letter 23 and the Letter 24 have been sent.

When: the dates of dispatch

A timeline chart is a type of chart that visually shows the chronological sequence of past or future events on a timescale. The chart’s timescale can be in days, weeks, months, quarters, years or other units of time depending on the type of events shown on it.

As explained in the previous section, the sending of a letter (an instance of pro:RoleInTime in the dataset) is characterized by having a series of properties. One of these properties is tvc:atTime, from the Time-indexed Value in Context ontology pattern (TVC), which relates the sending of a letter to the year/interval of years when such letter has been sent (its date of dispatch). In the dataset, each date of dispatch has been modelled as an instance of ti:TimeInterval, a class of the Time Interval ontology pattern (TI).

In the example below, the timeline chart shows the date of dispatch of each letter as either a dot (if it is an event defined in an instant) or a bar (if it is an interval between a given date and another, representing an uncertain attribution). Overall, the timeline consists of two bands: the upper band shows the timeline items with data within the selected timeline interval, while the lower band shows the original distribution of the events in the timeline. By clicking and dragging on the lower band it is possible to create a sub chart for zoom and selection range functionalities.

Who: the correspondents and the cited people

The letters corpus allows getting a more comprehensive look on the relationships between Vespasiano and his correspondents, other important humanists and nobles of his time.

A correspondent is a person who sent one or more letters to Vespasiano. As explained in the previous sections, the sending of a letter is an instance of pro:RoleInTime characterized by having a number of different properties. In addition to pro:relatesTo (which relates the sending of a letter to the exemplar of that letter), another property is pro:isHeldBy and relates the sending of a letter to the correspondent. In the dataset, a person is modelled as an instance of foaf:Person, a class which belongs to the Friend Of A Friend vocabulary (FOAF).

In the example below, another bar chart has been used to visualize each correspondent in relation to the number of letter he or she exchanged with Vespasiano. The horizontal axis shows the number of letters that has been exchanged between Vespasiano and his correspondents. The vertical axis shows the correspondents, each identified by his or her full name.

An arc diagram is a type of network graph in which the vertices are placed along a line, with edges being drawn as semicircles in one of the two halfplanes bound by it. An arc diagram is especially useful for visualizing ordered data.

A cited person is an instance of foaf:Person that is related to a piece of text, referencing that person and modelled as an instance of doco:TextChunk with type "attested form", through the property dcterms:references. In turn, that piece of text is related to its respective letter through the property dcterms:source.

In the example below, the arc diagram illustrates both the four most cited people in the letters (represented as teal-colored nodes) and the letters in which they are cited (represented as yellow-colored nodes) as well as the way in which they are related. This configuration allows one to clearly and easily observe all the letters in which a given person has been cited and, conversely, all the people that have been cited in a given letter.