Add disassociability use case: Privacy-Preserving Digital Advertising Measurement

[tina77fritz]

This PR adds a small reference tool demonstrating two foundational
disassociability techniques commonly used in privacy-preserving
analytics systems:

1. Contribution bounding – limiting how many times a single user may
   contribute to a cohort.
2. Cohort thresholding – reporting only cohorts that meet a minimum
   k number of bounded contributions.

What is included

• cohort_thresholding.py
  A simple, readable Python implementation using only the standard
  library. It loads an event-level CSV (user_id, cohort_id, value),
  applies contribution bounding and k-thresholding, and produces
  aggregated cohort-level metrics.

• example_events.csv
  A small synthetic dataset illustrating how the tool works.

• README.md
  Documentation covering:

  • the privacy motivation,
  • input/output formats,
  • example commands,
  • and notes on how the techniques support disassociability.

Purpose

This tool is intended as an accessible reference for practitioners who
need simple building blocks for privacy-aware aggregation workflows,
such as advertising measurement, recommendation systems, and other
event-based analytics pipelines.

The implementation is designed to be easy to read, adapt, and extend
(e.g., with differential privacy, synthetic data, or additional
metrics).

No production code is introduced; this is a documentation-and-example
tool aligned with the goals of the NIST Privacy Engineering
Collaboration Space.