Tracing the deportation to define Holocaust geometries. The exploratory case of Milan

Giovanni Pietro Vitali, Laura BrazzoCOLING 2024

TL;DR

This paper presents a pilot project that uses data visualization to map the historical routes of Jewish deportations from Italy during WWII. By combining archival data with interactive maps and network graphs, the project aims to provide a more detailed and accessible understanding of this tragic period, bridging collective memory with individual experiences.

What Problem Does This Paper Address?

The paper tackles the challenge of understanding and visualizing the complex historical dynamics of Jewish arrests and deportations from Italy between 1943 and 1945. A significant hurdle is the near-complete destruction of transport lists, making it difficult to reconstruct the precise movements of individuals. Existing research, while valuable, often lacks the granular, interconnected detail that digital visualization can provide. The authors aim to bridge this gap by creating interactive tools that allow users to explore both the broad patterns of deportation and the individual journeys of victims, thereby enhancing scholarly research and educational initiatives.

Methods

The project employs a multidisciplinary approach, combining archival research with advanced data visualization techniques. Key methods include:

  • Data Reworking and Structuring: Archival data from the Fondazione CDEC (Contemporary Jewish Documentation Center) is meticulously reworked and structured into data visualization models.
  • Data Visualization Models: The core of the project involves creating interactive maps, graphs, and network visualizations. These tools are designed using open-source technologies, primarily JavaScript libraries like Leaflet.js for maps and visNetwork for networks, along with R for data analysis and plotting.
  • Spatial Humanities Approach: The project draws inspiration from the "Spatial Humanities" field, using geographical information systems (GIS) to represent historical events and relationships. This allows for a "distant reading" of the overall phenomenon and a "close reading" of individual cases.
  • Interdisciplinary Collaboration: A team comprising a Digital Humanities expert, an archivist, a GIS developer, and an education manager collaborated to ensure both historical accuracy and technical innovation.

Datasets

The project utilizes data from the CDEC LOD (Linked Open Data) dataset of Holocaust victims in Italy. Specifically, two subsets of data were used:

  • Jews born in Milan (and arrested in Milan or elsewhere in Italy): 166 persons (reduced to 162 after exclusions).
  • Jews arrested in Milan (born in Milan or elsewhere in Italy or Europe): 278 persons (reduced to 259 after exclusions).

These datasets were acquired through SPARQL queries on the CDEC LOD endpoint, with criteria focusing on birthplace and place of arrest, respectively. The data was further enriched by incorporating information from archival handwritten paper cards, which often contain more granular details about locations.

Key Insights and Ideas

  • Bridging Collective and Individual Memory: The project successfully demonstrates how digital tools can connect the broad historical narrative of deportation with the specific, individual experiences of victims.
  • The Power of Interactive Visualization: Interactive maps and network graphs offer a more intuitive and engaging way to explore complex historical data than traditional methods, allowing users to discover relationships and patterns they might otherwise miss.
  • Data Gaps and Reconciliation: The project highlights the persistent challenges posed by incomplete historical data (e.g., missing dates, unknown perpetrators) but also showcases how archival work and careful data reconciliation can mitigate these issues.
  • "Distant" vs. "Close" Reading: The developed models enable both a broad overview of deportation dynamics ("distant reading") and a detailed examination of individual journeys ("close reading"), catering to diverse research needs.
  • Open-Source for Accessibility: The commitment to open-source technologies ensures that the tools and data are accessible for reuse by other researchers and educators.

Key Results

The project successfully created interactive visualizations, including:

  • Graphs: An interactive chronology of Jewish arrests in Milan (1943-1945), categorizing arrests by perpetrators (Fascist, Nazi, Unknown).
  • Maps: Two interactive maps visualizing deportation routes, arrest locations, and detention camps.
  • Networks: Interactive network graphs illustrating the journeys of deportees and the relationships between individuals and places.

These visualizations are publicly available on GitHub, serving as valuable research and educational resources. The accuracy analysis of the datasets revealed that while data concerning arrests in Milan is generally more precise, both datasets exhibit inaccuracies primarily related to dates.

Limitations and Future Work

  • Incomplete Data: The project acknowledges that the data, particularly concerning events before arrest, remains incomplete, limiting the ability to trace the entire pathway from birthplace to deportation site for all individuals.
  • Perpetrator Identification: A significant number of arrests in Milan lack identified perpetrators, posing a challenge for detailed analysis.
  • Future Objectives:
    • Establish a direct linkage between the visualization tools and the CDEC SPARQL endpoint to ensure data consistency.
  • Apply the developed model to the entire dataset of Holocaust victims in Italy.
  • Create an English-language version of the visualizations to expand their reach.

Summary updated

BibTeX
@inproceedings{vitali2024tracing,
  author = {Giovanni Pietro Vitali and Laura Brazzo},
  title = {{Tracing the deportation to define Holocaust geometries. The exploratory case of Milan}},
  booktitle = {International Conference on Computational Linguistics},
  year = {2024},
}
Figure