This code corresponds to the paper: Schäler, C., Hütter, T., & Schäler, M. (2023). Benchmarking the Utility of w-Event Differential Privacy Mechanisms-When Baselines Become Mighty Competitors. Proceedings of the VLDB Endowment, 16(8), 1830-1842.

The following list of software is required to execute the experiments:

• bash
• python 3
• curl
• java jdk
• java jre
• docker
• excel (for ploting)

We provide a convenience script to perform the experiments and plot the results shown in Schäler et al. (2023). Note that we assume that all datasets are stored in ./data. However, due to conflicting licenses, we are not allowed to share the experimental data publicly. The convenience script is executed from the root directory of the repository as follows:

sh scripts/perform-experiment.sh

After running the experiments, the aggregated raw data for Figures 2a, 2b, 3, 5 and 6, as well as Table 6 are located in the ./figures directory. To plot the data, open the vldb reproducability.xlsx file, also located in ./figures directory. It contains a tab for each .tsv file following the same naming convention

Further note that the experiments take approximately 24 hours to complete.

If you use this code for your work, please cite our paper:

@article{schaeler2023dpbench,
  author = {Sch\"{a}ler, Christine and H\"{u}tter, Thomas and Sch\"{a}ler, Martin},
  title = {Benchmarking the Utility of w-Event Differential Privacy Mechanisms - When Baselines Become Mighty Competitors},
  year = {2023},
  issue_date = {April 2023},
  publisher = {VLDB Endowment},
  volume = {16},
  number = {8},
  issn = {2150-8097},
  url = {https://doi.org/10.14778/3594512.3594515},
  doi = {10.14778/3594512.3594515},
  journal = {Proc. VLDB Endow.},
  month = {apr},
  pages = {1830-1842},
  numpages = {13}
}

Visit also our project website: https://dbresearch.uni-salzburg.at/projects/dpbench/index.html.