Showing posts with label data visualisation. Show all posts
Showing posts with label data visualisation. Show all posts

Jan 16, 2019

Visualizing the Art Market Networks of Hermann Voss

Network visualisation of Hermann Voss' connections, with his direct connections and the connection of his direct connections, according to the ALIU Final Report Red Flag list of Names of 1946.
(network graph generated automatically by google fusion tables-

With data visualisation tools, the graphical representation of the links between individuals in vast criminal networks is made possible. 

In the above graph, one sees the connections of Hermann Voss, as described in the 1946 OSS Art Looting Investigation Unit Final Report, which detailed the Nazi era art market network.

For each person in the ALIU "Red Flag List of Names" a brief paragraph summarises what is known about the person's role in the European Nazi-era art market. There are more than 1000 names on the list, each one with a few sentences of description, which typically includes a dozen or more attributes and connections.

The paragraph devoted to Voss in the Final Report contains 50 words, including his name.

Voss, Dr Hermann. Munich. Director of the Linz Special Commission, the Linz Musuem, and the Dresden Gallery from May 1943. Involved in the Schloss and Mannheimer collection (forced) sales, and the official chiefly responsible for Hitler’s looting and purchasing policies after 1945. In custody US 3rd Army, Munich, September 1945.

One can imagine that, due to practical contraints, the Art Looting Investigation Unit may have found it impossible to stuff everything they had learned about Hermann Voss into the one small paragraph allocated to each subject in the ALIU Final Report.

However, the "Voss, Dr Hermann" paragraph is not the only mention of Voss in the ALIU Final Report. 

A simple data filter in Sheets gives us the names linked to Voss in the ALIU Red Flag List/

Filtering for mentions of "Voss" in the shared Google Sheet "OSS ALIU Red Flag Names PUBLIC"

Voss, we see, is mentioned in 17 entries in the ALIU Red Flag List*.

Goepel, Grosshennig, Gurlitt [the father of Bavarian art hoarder, Cornelius], Nadolle, Oertel, Pat-Zaade, Posse, Reimer, Schilling, Schmidt, Voss, Waldner, Weber, Zinckgraf, Mandl, de Boer, Hoogendijk,.

Each one of the above individuals, who the ALIU believed to be connected to Voss, had his own network, or "cluster".  
It is by joining together these different clusters that we are able to extract information that can be used to visualize the Nazi art market network - as analysed by the Art Looting Investigation Unit in 1946.

It is these mentions that we will use to select the data for network analysis.


How to map the direct connections of Dr. Hermann Voss using Google Sheets and Google Fusion Tables

In our experiments we used the following method to prepare the data:

1) copy the entire ALIU Red Flag list into a Google Sheet called OSS ALIU Red Flag Names PUBLIC.  It is this file that we will return to, again and again, to find more information about different networks contained in the Final Report.  This task needs to be performed only once.
(This Google sheet can be viewed by anyone with this link.)

2) filter for Voss (Data, Filter by Condition, Text contains "Voss") as shown above.

3) copy the results (which are all the Red Flag texts in which Voss is mentioned) to another sheet, named, for example, "Voss Network"

4) Format the "Voss Network" sheet (so that it can be loaded and analysed in Google Fusion Tables)
Sheet AFTER formatting: Name1 and Name2 are linked in the network

Note: We format the data only after having gathered it all together. 

4.1 Using the Sheets Data function, Split Text to columns, Use the "." as the Separator.

4.2 Add a column to the left .

4.3 For all the entries that mentioned Voss, copy "Voss, Dr Hermann" into the new column.
This is important for network analysis later, because it links Voss to the name of the Red Flag entry in which his name appears.

Note: the assumption is that the mention of Voss' name implies some kind of link to Voss. The exact nature of the link is specified in the text, which is information that we will exploit later using a different network analysis tool, but which we do not need at this stage in the analysis.

4.4  Name the columns (from Left to Right): Name1, Name2, Location, Role1, Role2, Role3...Role10

(Note: We format all the Sheets the same so that later we can mix selected files in Google Fusion Tables)

5) Load the Google Sheet "Voss Network" into Google Fusion Tables

(Note: If you have multiple sheets (tabs) you can select which one to load)

6) In Google Fusion tables, click on "Create Chart"./ Specify that the chart is for Name1, Name2

note! This will generate a circle with spokes in which  "Voss, Dr. Hermann" is in the centre with links to every Red Flag Name entry that mentioned him.

Try it!

How to add clusters to the network

To obtain the graph with several clusters at the beginning of this post, one inserts a step between 3) and 4). in which one adds the connections for some or all of the names whose entries mentioned Voss.

Before formatting the "Voss Network" Google Sheet, one returns to the OSS ALIU Red Flag Names PUBLIC Google Sheet and filters for Red Flag entries that mentioned Voss. Goepel is the first name on the list,.

Filter for Goepel and copy the result into the "Voss Network" Google Sheet, and repeats for the next name.

ALIU entries that mentioned Goepel are: Goeple, Wuester, Holzapfel, leegenhoek, Lefranc, Mandl, Block, and Cramer.  We can think of this as the "Goepel" cluster.

This will add to the Voss network, all the Red Flag entries that mentioned Goepel.

Once all the information one wants about Voss' contacts has been added, proceed to formatting, as described in step 4, then load in Google Fusion Tables

Note: Attention, in the column to the left (that will become Name1), one puts Goepel (or whatever name was filtered for)

Direct Links to Voss and Direct links to one of Voss' direct links, Goepel.
Note that the names appear exactly as they do in their ALIU entry, "Voss, Dr Hermann" and "Gopel, Dr Erhard".
We are careful to use the exact ALIU names so that they will match in network analysis.

This approach can be used for any set of links one wants to analyse, whether to a name of a person, a place, an organisation, an event,  a year or any word that appears in the ALIU Red Flag list.

Data, always in the same format, can be mixed and matched, a process we will show later.

Links to Datasets:

Considerations concerning the source

By necessity, important information will be missing from an operational report made under time constraints in difficult conditions.  We are, in using the information in the ALIU Red Flag List to map our networks, trusting the selection of information made by the OSS ALIU team.

Is there - should there be- a limit to this trust? What are the contours to the report's reliability? Are there blindspots we should be aware of? Omissions that should be compensated for by bringing in other sources? Attention to some areas that seems excessive?

For the time being we will acknowledge these questions and put them aside, for further exploration later. No source is perfect, and, at present, the 1946 ALIU Red Flag List, with all its imperfections, still seems to be the most useful primary historical source we have.

*Vossisk was picked up but is not a contact. (In addition, Voss had an entire detailed interrogation report (DIR) dedicated to him, but we are not using the detailed report, only the summary of Red Flag names in the Final Report.)

Jan 15, 2018

Amazing Wikidata Timelines

Wikidata has timelines built into its queries. (See above for "People who died by burning")

First of all, this is incredible and amazing and insanely powerful.