• Intro
• Insights
  • Alfred Döblin: Berlin Alexanderplatz
  • Hans Fallada: Every man dies alone
• Technology
• Data
  • Data origin
• Summary

Intro

Jan-Erik Stange from the Urban Complexity Lab at FH Potsdam utilizes references of locations in novels from different times to create appealing visualizations.

One chart type, the “narrative view”, shows the locations (dots) horizontally through the progress of the book. Vertically, they are ordered by frequency of appearance. If different locations are mentioned in conjunctions, they get connected by vertical lines (Fig. 1).

Another chart type shows the locations as a geographic map. It represents the frequency of mentions by opacity. Locations not present in the book are invisible (Fig. 2).

Insights

Let’s take a look at two charts for two of the novels and examine, what we can find out in those visualizations:

Alfred Döblin: Berlin Alexanderplatz

I explored the narrative view, by taking a closer look at the particular section, where a lot of points appear vertically aligned over each other. This means, a lot of locations are mentioned close to each other. (Fig. 1)

This paragraph in the book (set in the 1920s) describes one of the many “pictures” in the book, which are not connected to the main storyline but create the unique atmosphere of the book. The paragraph describes an affair between a young girl and an older man. The many locations appear in an introducing sentence, where a part of the tram map is recited. The unconnected story of the young girl and her lover is somehow justified by the tram network, that appears to connect not only the different stations in the city but also the stories that are happing there. Long vertical lines which contain multiple dots emphasize such interconnections, that are present throughout the progress of the whole book.

Hans Fallada: Every man dies alone

The map view shows only locations, which are present in the novel. The opacity of the locations indicates the frequency of appearance. The layout is based on the geo-coordinates of the streets and places. Berlin’s river Spree is also visible. (Fig. 2)

An obvious characteristic of the locations shown on the map is a focus on two regions: The famous Alexanderplatz in the northeast of the river, and the surroundings of Nollendorfplatz in Berlin Schöneberg. Schöneberg is the part of the city, where the two main characters in the book, the married couple Anna and Otto Quangels, are living.

The region around Alexanderplatz shows up for multiple reasons: It’s the location of the police department, where the Gestapo is trying to hunt down the source of the mysterious anti-nazi-flyers, which are actually written by the Quangels. Also, the Alexanderplatz represents one of the most frequented places in Berlin. The Quangels are visiting it quite often, to deploy their flyers, but also visit bars, meet other people or for traveling through.

I guess, Alexanderplatz would show up in most novels set in Berlin in the early 20th century.

Technology

The interactive visualization is empowered by D3.js.

The map view is created by making use of d3.v3s geo-module to render preprocessed data in Mercator projection. This projection preserves the angles and the shapes of smaller objects, therefore it is a good choice for city maps. Online map services, e.g. Google Maps or OpenStreetMap, are utilizing it.

The narrative view draws the lines with d3 in a custom way.

To provide a transition between the two different charts, all data points in the first view are moved to the center and faded out. Then the next view is faded in with all data points in the center. They are then animated to reach their final positions.

Data

The three novels visualized are:

• Alfred Döblin: Berlin Alexanderplatz (set in the 1920s)
• Hans Fallada: Alone in Berlin (set in 1940-42)
• Sven Regener: Berlin Blues (set in 1980)

A single TSV file maps the mentions of locations in the books to an address string. A JSON file with geo information for the locations. A second JSON file provides all nodes, together with the information, which locations are mentioned together. A second TSV file contains the paragraphs of the novels with text and extracted nouns.

Data origin

It’s not explained which text-source has been used or how the locations have been extracted from the novels. It might be anything from a manually compiled list to an advanced Named Entity Recognition algorithm. The content of the fourth data file indicates, that the locations have been identified via noun extraction, maybe complemented by mapping with a list of locations.

However, two facts are interesting:

• The Alexanderplatz is correctly identified, even if it’s been referred as “Alex”. Alex is a common nickname of the place, but also a common forename. An automated algorithm probably requires beeing adapted, to differentiate between those.

• Sometimes obvious locations have not been added to the data: In the paragraph explored above (Fig. 1), “Seestraße” is mentioned (“[…] Seestraße Ecke Togostraße […]”). But no dot appears on the corresponding line in the narrative view. There might be a good reason for this, but it also would make perfect sense, if the demonstration of the visual interface just had a higher priority in the project, than the data quality.

Summary

As a technically interested person, I’m missing some information about the origin/compilation of the data. But in all other aspects, I really appreciate the Novel City Map, because I love new ideas on how to explore commonly known data like novels through a visual and data-driven approach. And I would love to discover other novels from other cities in the near future… :-)

[ via nerdcore.de ]