The Namescape Visualizer

At Namescape Visualizer the networks of fictional characters in the books in the Namescape corpus are displayed in various ways.

Searching

One can search by author name and book titles, possibly filtered by publication data.

Book display

The book display page contains the publication metadata, and (if found on the internet), an image of the cover.
The twenty characters with the highest dispersion value are listed. NB: due to a sometimes rather infelicitous attempt to group variant designations of the same name (Jan, Jan Janssen), the displayed normalized form is sometimes garbled.

Each character is (if possible) linked to a corresponding entry in Dutch wikipedia.

Visualisations

For each book, the network of characters is visualized in three ways: two different representations of the cooccurrence network, and a dispersion plot.

Network of characters

Each book contains a network of characters. Two characters are considered connected if they are both mentioned in the same paragraph.

In the resulting graph, strongly interconnected clusters of the network can be identified. The character bundle and the matrix graph are different ways of displaying the network.

The user can order the characters by name, frequency or by cluster. The clustering has been performed according to the Louvain method. The colors in the matrix graph correspond to the clusters, the intensity of the color indicates frequency of the name in the book.

Dispersion graph (barcode graph)

The dispersion graph shows which character is mentioned in which paragraph. The characters in the plot are sorted (in descending order) by dispersion. The horizontal axis represents paragraphs in book order, the vertical axis corresponds to characters. A colored bar at (x,y) means that character y is mentioned in paragraph x. Dispersion seems a good indicator to track the main characters in a novel. Dispersion plots are not unlike the more artistic narrative charts.

Sorting and filtering

The user can choose the criterion by which characters are sorted, and which portion of the characters should be included in the plots.Possible filters are degree (number of characters to which a character is connected in the network) and dispersion.

.