This project evaluates the impact of a major redesign of Vanguard’s client-facing digital interface, using an A/B test (Mar–Jun 2017). The goal is to determine if the new design improves user experience and process completion rates by at least 5%, as measured by user behavior and demographics.

• The Control Group used Vanguard’s traditional interface.
• The Test Group used the redesigned UI with enhanced prompts and cues.

• Google Slides Final Presentation:
  Project Slides
• Tableau Dashboard:
  Interactive Visualizations & KPIs
• Repository:
  Full documented code, data integration, analysis workflow, and methodology.

• df_final_experiment_clients: Assigns group (Test/Control) to each client.
• df_final_web_data: Logs each client’s process steps and timestamps.
• Client demographics dataset: Age, gender, tenure, account balances, engagement (calls, logons), etc.

1. Data Integration & Cleaning:
   All datasets were merged, cleaned, and prepared for analysis using Python.
2. Analysis:
   Explored client segmentation by age, tenure, and engagement, and measured key metrics:
   • Conversion Rate: % of users completing the process.
   • Error Rate: Frequency of backward navigation (proxy for confusion/errors).
   • Step Time: Average dwell time per process step.
3. Experiment Results:
   • The new UI led to 31% fewer user errors (backward steps).
   • Middle-aged (30–50) and senior clients are the most active users; seniors show highest average balances.
   • Male clients average almost double the account balances of females in this sample.

• notebooks/: Main Jupyter Notebooks for data exploration, cleaning, analysis, and visualization.
• data/: (Not versioned) Place for raw, cleaned, or derived datasets.
• figures/: Generated plots, charts, and figures for the project and presentation.
• sql_scripts/: Supplementary SQL queries for advanced data manipulation.
• src/: Custom scripts and modules (if any) used throughout the analysis.
• config.yaml: Analysis workflow configuration parameters.
• pyproject.toml, uv.lock: Project/dependency management for a reproducible Python environment.
• README.md: Full documentation with detailed workflow, context, and observations.
• README_experience.md: This file — summary, outputs, repository structure.

• The redesigned user interface reduced errors and improved process flow.
• Seniors have the highest balances and robust platform usage.
• All results, code, and dashboards are available for further stakeholder review.