Cardiovascular Biometrics: A Systematic Data Audit & Inferential Analysis

1. Executive Summary

This technical audit targets a dataset of 1,000 patient records to validate data integrity and assess the impact of lifestyle interventions on cardiovascular performance. The analysis identifies critical biometric inconsistencies and employs a frequentist statistical approach to test hypothesis significance.

2. Data Integrity & Risk Mitigation (GIGO Prevention)

A preliminary audit revealed structural flaws in the source data that would lead to "Garbage In, Garbage Out" (GIGO) scenarios.

• Audit of Faulty BMI: The original BMI column exhibited mathematical drift. I implemented a recovery field BMI_Real derived from raw Weight_kg and Height_cm.
• Structural Deconstruction: The Blood_Pressure string variable was parsed into Systolic and Diastolic integers, enabling granular clinical risk profiling.
• Validation Metric: A Pearson correlation matrix confirms the audit's success, showing a robust 0.82 correlation between weight and the corrected BMI.

3. Inferential Framework & Hypothesis Testing

The core objective was to determine if physical activity levels significantly shift the biometric mean of heart rate.

• Statistical Methodology: Two-sample Student's T-test.
• Null Hypothesis ($H_0$): No significant difference exists in heart rate between active and sedentary cohorts.
• The Verdict: Computed P-Value = 0.4614.
• Strategic Insight: We fail to reject the null hypothesis. The variance is not statistically significant, proving that cardiovascular health in this cohort is a multifactorial system.

4. Technical Stack

• Environment: JupyterLab.
• Language: Python 3.10+.
• Libraries: Pandas (Auditing), Seaborn/Matplotlib (Visualization), Scipy.stats (Inferential Analysis).

---

Audit performed by André Vinagre.