"Hidden in Plain Bytes" Source Code

This repository contains code and documentation for processing data exports from major online platforms, used in the research paper:

Julia Nonnenkamp, Naman Gupta, Abhimanyu Dev Gupta, and Rahul Chatterjee. 2025. Hidden in Plain Bytes: Investigating Interpersonal Account Compromise with Data Exports. In Proceedings of the 2025 ACM SIGSAC Conference on Computer and Communications Security (CCS '25). ACM, Taipei, Taiwan, 1–14 (October 13–17, 2025). DOI: 10.1145/3719027.3765147.

Overview

This repository processes 12 data exports obtained under "right of access" requests (GDPR/CCPA) from 6 major online platforms: Apple, Discord, Facebook, Google, Instagram, and Snapchat. The data was collected to investigate and identify interpersonal account compromise patterns using real-world data exports.

Dataset Access

The processed dataset is available on Zenodo with controlled access due to the sensitive nature of Tech-Facilitated Abuse (TFA) research. Access is restricted to prevent misuse.

To access the data:

1. Request access via the Zenodo repository
2. Download the dataset files: sam_january_cleaned.zip and sam_february_cleaned.zip
3. Extract both ZIP files recursively in your working directory before running the transform.zip notebook.

Environment Setup

Prerequisites

• Python >=3.12
• Git

Installation

1. Clone this repository:

   git clone https://github.com/WISPR-lab/data-exports-tfa.git
   cd data-exports-tfa

2. Create and activate a virtual environment:

   python -m venv .venv313
   source .venv313/bin/activate  # on Windows: .venv313\Scripts\activate

3. Install dependencies:

   pip install -r requirements.txt

Usage

Quick Start

1. Prepare your data: Ensure the dataset ZIP files are extracted in your working directory
2. Open the analysis notebook: Launch transform.ipynb in Jupyter or VS Code
3. Configure paths: Follow the instructions in the notebook to set your data input and output paths
4. Run the analysis: Execute the notebook cells to reproduce the data transformation and analysis

Notebook Structure

The main analysis is contained in transform.ipynb, which:

• Parses raw data exports from multiple platforms
• Transforms nested JSON/CSV structures into flattened DataFrames
• Groups similar data elements across platforms according to characteristic features (see Section 3 or )

Supporting Scripts

The scripts/ directory contains modular utilities:

• parse.py: Core parsing and data transformation logic
• group_utils.py: Functions for clustering and grouping similar data elements
• timeutils.py: Time-related parsing utilities

For documentation of parser utilities and data structures in an extended appendix, see appendix.md.

License

• Code: Released under the MIT License
• Paper: Protected under Creative Commons Non-Commercial ShareAlike license
• Dataset: Available under CC International license with controlled access

Acknowledgments

This research was conducted at the University of Wisconsin–Madison.