Introduction
The diamond industry in both Belgium and the Netherlands was dominated by Jews. During the Middle Ages, crafts and trade were dominated by guilds that excluded Jewish labor from participation, but no such guild existed for the craft of cutting and polishing diamonds. This meant that the diamond industry was historically one of the few industries open to Jews. During the heyday of the diamond Industry in the Netherlands, which was concentrated in Amsterdam, almost all of Amsterdam’s Jewish families were involved in the trade one way or another. The Amsterdam diamond industry declined during the First World War, but only the murder of most of Amsterdam’s Jews and the concomitant devastation of the diamond trade during the Second World War marked the definitive end of Amsterdam’s influential position in the diamond trade. It was Antwerp that took over Amsterdam’s role as one of the dominant forces in the global diamond market (Laureys 2005; Joods Cultureel Kwartier 2019).
Scholarly work on the Netherlands and the Holocaust shows that three-quarters of the Dutch jews were murdered during the Holocaust, most of them in Sobibor and Auschwitz (van der Boom 2012). In the case of Amsterdam, between 74.3 and 75.3 percent of all the Jewish residents were murdered (Tammes 2017). Men were more likely to be selected to perform forced labor in concentration, forced labor and extermination camps than women, who were more likely to be sent to death camps to be murdered soon after arrival. While this clearly mattered for the gender distribution of Jews regarding the camps they ended up in, Tammes finds that in the case of the Amsterdam Jews, gender differences did not matter much for the survival rate (Tammes 2017).
The intimate connection between the fate of the diamond industry in the low countries and the fate of the involved Jews during the Second World War can be demonstrated and mapped out by using Linked Data. Information about the lives of a large number of Dutch (and also some Belgium) victims of the Holocaust comes from two online databases, namely the IISG ANDB and ADB database and the Oorlogslevens database (IISG 2019). The databases provide information on matters such as when and where people were born, their parentage, where they lived, how they developed professionally within the diamond trade and the General Diamond Workers’ Union of the Netherlands (ANDB) and the General Diamond Workers’ Association of Belgium (ADB) and on when and where people died (IISG 2019; Oorlogsbronnen 2021). Combining such data allows for storytelling exercises about the lives of Jewish victims of the Holocaust who were at some point involved in the diamond industry and the ANDB and ADB. These stories are essential to keep alive the memory of the horrific fate of the Jews during the Second World War, but also to highlight their great societal achievements and the potential squandered in their murder. That such fine-grained data is available in the first place is due to the professional organization of the unions, which involved well-organized member records, but also to the prominent role played by the ANDB in Dutch trade union history. The fact that the ANDB was the first modern trade union in the Netherlands and achieved significant accomplishments, including the eight-hour working day, benefits in the event of sickness and unemployment and equal wages for males and females, prompted the ISSG to digitize the ANDB’s extensive archive in order to bring its history and achievements closer to the general public (IISG 2019).
In this short data story we ask how Linked Data can help map the lives and deaths of holocaust victims and what the limits of Linked Data are to that mapping. More specifically, we focus on victims who were at some point in life members of the ANDB and/or ADB. Until now, this link between the ANDB and ADB data and the data provided by Oorlogslevens has not yet been made or queried on druid.datalagend.net. Our data story is somewhat funnel shaped, where each element is visualized, discussed and subjected to source criticism. We start by providing the most comprehensive information available for our purpose, namely information about where the victims died. In order to do this we have cleaned the underlying assets with regard to place of death to ensure uniformity and removed al non-person entries in R. We zoom in on gender to see whether the earlier mentioned influence of gender on deportation destination is also reflected in the data provided by the ISSG and Oorlogsverhalen. We then continue with a much smaller sample of people of which the place of birth could also be traced. We subsequently map what the role of these people was in the diamond industry, and who of this small sample lived in Amsterdam and in which street both during their lives and before deportation. We then show the ANDB and ADB membership cards of the individual members of our limited sample to bring the reader closer to the primary sources underlying the mapping of the lives of these individuals. We end our data story with a short conclusion in which we answer our research question.
Analysis
Using the yet unlinked Oorlogsbronnen dataset containing just the names of ANDB-ADB members, we were able to create a linked dataset containing unique person identifiers (UUIDs), names, gender, dates of death, and - most importantly for us - places of death. The table below contains all individuals of which we know the name, the gender, and the place of death. It is important to note here that on top of issues with spelling variations for places of death corrected by hand, and incorrectly entered data carefully filtered out with an R script, the place of death was not recorded for everyone in the Oorlogsbronnen dataset. We are also aware of the fact that with the standardization of the places of death, more fine-grained information, for example about which specific camp of Auschwitz people died, is lost. However, to properly outline the big picture, we have chosen to make this trade-off. Hence, the following table already presents a selection of 5,276 people out of 5,440. We left out the date of deaths of these people, because these were entered with varying standards, but in theory these could be added to the list, too, to check when people most often died at which place.
Deathplace List Query
This list of ANDB or ADB members with their gender and place of death may be neatly compiled into figures through the use of the COUNT function as exemplified below. The first pie-chart shows the distribution of death places for the 10 most common death places regardless of gender while the second pie-chart takes account of this variable. That extermination camps Auschwitz and Sobibor make up the vast majority of death places may hardly come as a surprise, but the chart taking account of gender does present some interesting findings. While 97.5 percent of the female members of the ANDB and ADB persecuted by the Nazis died in Auschwitz and Sobibor, this figure was 87 percent for men with rather vague places of death such as Central Europe, Poland and unknown places among the 10 most common places of death. Especially Central Europe as a place of death takes a substantial claim with 5.5 percent of the total number of deaths of male members of the ANDB-ADB. Although not entirely pointing towards an unknown place of death, statements such as Central Europe and Poland cannot lead to an accurate timeline of an individual's life. At the same time, linked data poses the opportunity to automatically update these figures once a new dataset with more accurate data is linked using the same vocabularies.
The table below shows the names, gender, birth places and death places of those ANDB and ADB members of whom we know all these four different variables. This immediately results in a (reverse) decimation with less than 10 percent of the first subselection remaining. However crude, this small selection will allow us to trace these individuals' lives from their birth to their death both in space and time.
The map below shows another subselection of the ANDB Holocaust victims. This map shows the addresses of the Amsterdam members mentioned in the table above that had their addresses registered by the ANDB on their membership card(s). It should be stressed that the numbers on the map do not represent unique individuals, but registered membership cards. If an individual had multiple membership cards, that individual is therefore represented more than once. Though it cannot be claimed with certainty that these were the final addresses inhabited by each of the victims, the map highlights a general spatial concentration of Jewish diamond workers' residencies in the south-east part of the city. This concentration corresponds with the infamous 1941 Amsterdam "stippenkaart" in which Dutch public officials plotted the spatial distribution of the Jewish city population on an Amsterdam city map for the Nazi-German occupiers. That map, quite similarly, highlighted a concentration of the Jewish population in the south-east part of the city.
The membership cards of the ANDB in addition to spatial information also allow to trace information on the (professional) lives of these victims. Apprenticeship cards and membership cards registered occupations within the diamond industry, marriage partners and children. From the subselection of individuals for whom a place of death was registered in the Oorlogsleven database, their membership cards, full name, and (if available) occupation within the diamond industry, birthplace, birthdate, place of death and date of death can be queried below. A major advantage of Linked Data is that when scanned and inserted into the online database, you can check primary source material yourself. Unfortunately, in our case this is only possible for the ANDB data. Records of place of death have not been scanned and can therefore not be displayed or checked. This again highlights that the opportunities of Linked Data are ultimately bounded by the way in which the original sources have been converted into an online data set.
Conclusion
By filtering person-level data with increasing layers of variables and showing the intermediate steps, we have shown that it is possible to trace the lives of historical individuals by matching linked data sets. These individual level details, as shown in the charts and map above, can also be quite effortlessly aggregated to present macro analyses without creating new datasets or transforming the underlying dataset. However, it is quite clear that despite the possibility to add data to the analysis without even having to change anything regarding the data story or its underlying queries, because it would update automatically, the underlying datasets do require enough standardisation and sufficient primary source criticism on behalf of those who uploaded it. This poses a threat to the integrity of the the quantitative research cycle, as it may be hard to trace linked data back to the original .csv file and the primary source on which that depends. However, if one has taken account of these issues and is aware of how to overcome them, the possibilities of linked data may be endless.
Bibliography
Laureys, E. 2005. Meesters van het diamant: de Belgische diamantsector tijdens het nazibewind. Tielt: Lannoo.
Tammes, P. 2017. “Surviving the Holocaust: Socio-demographic Differences Among Amsterdam Jews.” European Journal of Population 33 (3): 293-318.
Van der Boom, B. 2012. 'Wij weten niets van hun lot'. Gewone Nederlanders en de Holocaust. Amsterdam: Boom Amsterdam.
Websites:
Joods Cultureel Kwartier: https://jck.nl/en/node/4040