---

name: ds-data-privacy compatibility: opencode completeness: 95 content-types:

• code
• guidance
• do-dont
• examples description: '"Applies privacy-preserving techniques including anonymization, differential privacy, encryption, and GDPR compliance for sensitive data"' license: MIT maturity: stable metadata: domain: coding output-format: code related-skills: ds-data-versioning, ds-privacy-ml role: implementation scope: implementation triggers: data privacy, anonymization, differential privacy, GDPR, PII protection privacy-preserving, sensitive data archetypes:
  • tactical
  • generation anti_triggers:
  • brainstorming
  • vague ideation
  • code golf
  • over-engineering response_profile: verbosity: low directive_strength: high abstraction_level: operational version: "1.0.0"

---

Data Privacy

Comprehensive guide to data privacy in machine learning and data science workflows.

When to Use This Skill

• Solving real-world data collection & ingestion problems
• Building machine learning pipelines with data privacy
• Implementing best practices for data privacy
• Optimizing model performance using data privacy techniques
• Learning industry-standard approaches to data privacy

When NOT to Use This Skill

• When using pre-built libraries without understanding underlying concepts
• For toy problems that don't require data privacy rigor
• When domain expertise in specific problem requires different approach
• If your problem doesn't require the complexity this skill provides

Purpose and Key Concepts

Data Privacy is a critical component of the machine learning workflow. This skill covers:

1. Theoretical foundations — Mathematical principles and statistical concepts
2. Practical implementation — Working code examples and patterns
3. Common pitfalls — Mistakes to avoid and how to recover from them
4. Best practices — Industry-standard approaches and optimization techniques

Core Workflow

1. Understand the problem — Clearly define what you're solving for
2. Select approach — Choose the right technique for your data and constraints
3. Implement solution — Write clean, tested code following best practices
4. Validate results — Verify your implementation with tests and validation
5. Optimize performance — Improve efficiency and accuracy incrementally

Implementation Patterns

Pattern 1: Basic Data Privacy

import pandas as pd
import numpy as np
from typing import Dict, Any

TOKEN_SPACE_SIZE: int = 10000

def apply_basic_privacy(data: pd.DataFrame, epsilon: float = 1.0) -> Dict[str, Any]:
    """
    Apply basic privacy-preserving transformations.
    Implements tokenization for categorical PII and Laplace noise for numerical data.
    """
    if data.empty:
        raise ValueError("Input DataFrame cannot be empty")

    protected_data = data.copy()
    metrics: Dict[str, int] = {"columns_anonymized": 0, "noise_added": 0}

    for col in protected_data.columns:
        if protected_data[col].dtype == "object":
            protected_data[col] = protected_data[col].apply(
                lambda x: f"token_{abs(hash(str(x))) % TOKEN_SPACE_SIZE}" if pd.notna(x) else x
            )
            metrics["columns_anonymized"] += 1
        elif np.issubdtype(protected_data[col].dtype, np.number):
            sensitivity: float = protected_data[col].max() - protected_data[col].min()
            scale: float = sensitivity / epsilon
            noise: np.ndarray = np.random.laplace(0, scale, size=protected_data[col].shape)
            protected_data[col] = protected_data[col] + noise
            metrics["noise_added"] += 1

    return {"protected_data": protected_data, "metrics": metrics}

Pattern 2: Production-Ready Data Privacy

import logging
import pandas as pd
import numpy as np
from typing import Any, Dict, List

logger = logging.getLogger(__name__)

class DataPrivacyEngine:
    """Production-grade privacy engine implementing GDPR-compliant anonymization."""

    def __init__(self, epsilon: float = 1.0, k_anonymity: int = 5) -> None:
        self.epsilon: float = epsilon
        self.k_anonymity: int = k_anonymity
        self._log: logging.Logger = logging.getLogger(self.__class__.__name__)

    def execute(self, data: pd.DataFrame, sensitive_cols: List[str]) -> Dict[str, Any]:
        """Execute privacy transformations on sensitive columns."""
        if not isinstance(data, pd.DataFrame):
            raise TypeError("Expected pandas DataFrame")
        if data.empty:
            raise ValueError("Input data cannot be empty")

        protected: pd.DataFrame = data.copy()
        results: Dict[str, Any] = {"status": "success", "transformations_applied": []}

        for col in sensitive_cols:
            if col not in protected.columns:
                self._log.warning(f"Column {col} not found, skipping.")
                continue

            if protected[col].dtype == "object":
                protected[col] = protected[col].apply(
                    lambda x: f"cat_{abs(hash(str(x))) % 100}" if pd.notna(x) else x
                )
                results["transformations_applied"].append(f"tokenized_{col}")
            else:
                col_min: float = protected[col].min()
                col_max: float = protected[col].max()
                sensitivity: float = col_max - col_min
                scale: float = sensitivity / self.epsilon
                noise: np.ndarray = np.random.laplace(0, scale, size=len(protected))
                protected[col] = protected[col] + noise
                results["transformations_applied"].append(f"dp_noised_{col}")

        results["protected_data"] = protected
        self._log.info(f"Privacy engine completed. Applied {len(results['transformations_applied'])} transformations.")
        return results

Best Practices

• ✅ Always validate your implementation on test data
• ✅ Document your assumptions and methodology
• ✅ Use version control for reproducibility
• ✅ Monitor performance metrics in production
• ✅ Periodically review and update your approach
• ✅ Test with edge cases and outliers
• ✅ Log all significant operations for debugging

Common Pitfalls

| Pitfall | Problem | Solution | |

---

---

Constraints

MUST DO

• Validate all data preprocessing steps are fit-only on training data, never on validation or test sets
• Implement reproducible pipelines with fixed random seeds and deterministic operations where possible
• Report model performance with confidence intervals via bootstrapping or cross-validation across multiple runs
• Log all experiments with parameters, metrics, and artifacts using MLflow or equivalent tracking system

MUST NOT DO

• Do not evaluate a model on the same data used for training — always hold out a proper test set
• Avoid overfitting to the validation set by limiting hyperparameter search iterations
• Never use features that can only be computed at inference time (look-ahead bias)
• Do not report single-run accuracy without statistical significance testing or error bars

Live References

Authoritative documentation links for this skill's domain. The model follows markdown links at load time to resolve external references and inline content.

• Data Privacy — Wikipedia
• NIST Privacy Framework
• GDPR Official Text (europa.eu)
• HIPAA Privacy Rule (HHS)
• Data Protection Impact Assessment Guide (EDPB)