Skip to content
 
 

Repository files navigation

🧩 Vanguard Digital Experiment — Module 2 Project

📘 Introduction

Welcome to the Module 2 Project!
In this project, you will apply your skills in data analysis, hypothesis testing, and visualization to evaluate the results of a digital experiment.
You will perform EDA, performance analysis, hypothesis testing, experiment evaluation, and data visualization in Tableau.


🏢 Project Context

You are a Data Analyst in the Customer Experience (CX) team at Vanguard, a U.S.-based investment management company.

Your first assignment: analyze the results of a digital experiment designed to improve client experience.

🎯 The Digital Challenge

Vanguard wanted to test whether a new, modern User Interface (UI) — with in-context prompts (messages, hints, or cues) — could encourage more clients to complete an online process compared to the traditional design.


🧪 The Experiment

An A/B test was conducted to compare client behavior between two groups:

  • Control Group: Clients using the traditional online process
  • Test Group: Clients using the new UI design with in-context prompts

Both groups followed the same flow:

  1. Initial page
  2. Three process steps
  3. Confirmation page (completion)

The main goal:
➡️ Determine if the new design improves completion rates and user experience.


📊 Datasets

You will work with three datasets:

Dataset Description
df_final_demo Client demographic data (age, gender, account details, etc.)
df_final_web_data_pt_1 & df_final_web_data_pt_2 Digital activity logs (should be merged before analysis)
df_final_experiment_clients Information on which clients participated in the A/B test

🧱 Metadata

These are the key variables used in the project:

Column Description
client_id Unique client ID
variation Indicates experiment participation (control/test)
visitor_id Unique client-device combination
visit_id Unique session identifier
process_step The digital process step visited
date_time Timestamp of web activity
clnt_tenure_yr Client tenure in years
clnt_tenure_mnth Client tenure in months
clnt_age Client’s age
gendr Client’s gender
num_accts Number of accounts held
bal Total account balance
calls_6_mnth Number of support calls in the last six months
logons_6_mnth Number of logins in the last six months

🗓️ Project Structure

1. EDA & Data Cleaning

  • Explore all datasets using Pandas, Matplotlib, and Seaborn
  • Identify and fix data quality issues
  • Perform client demographic and behavior analysis
    • Identify key client segments (age, tenure, balance, etc.)
    • Explore behavioral trends and process interactions

2. Performance Metrics

Define and compute key performance indicators (KPIs):

  • Completion Rate: % of users reaching the final confirmation step
  • Time Spent per Step: Average duration between steps
  • Error Rate: Frequency of users moving backward in the process

Compare the performance of the new design vs. traditional design using these KPIs.


3. Hypothesis Testing

Perform statistical testing to evaluate the redesign’s effectiveness:

  • Test differences in completion rates between groups
  • Assess whether improvements are statistically significant
  • Optionally evaluate cost-effectiveness thresholds

4. Experiment Evaluation

Evaluate the experiment from a design perspective:

  • Was the design fair and balanced between groups?
  • Was the duration sufficient for reliable results?
  • What additional data could improve confidence in the findings?

5. Visualization (Tableau)

Create an interactive Tableau dashboard that displays:

  • Completion rates and group comparison
  • Demographic distribution
  • Time spent and error rates
  • Key insights and experiment outcomes

6. Final Deliverables

  • Cleaned datasets
  • EDA and insights notebook
  • KPI analysis and hypothesis testing results
  • Experiment evaluation summary
  • Interactive Tableau dashboard
  • Final presentation and report

🧰 Tools & Technologies

  • Python: pandas, numpy, matplotlib, seaborn, scipy
  • Tableau: for interactive dashboards and storytelling
  • Kanban board: for task management (Trello, Notion, or Jira)
  • Optional: Streamlit for additional dashboard interactivity

👥 Collaboration

This is a pair project.
Use a Kanban board to divide tasks efficiently and track progress.
Suggested role split:

  • Analyst 1: EDA, Data Cleaning, KPI Analysis
  • Analyst 2: Hypothesis Testing, Experiment Evaluation, Tableau Visualization

🎓 Learning Outcomes

By completing this project, you will:

  • Apply EDA and data cleaning techniques
  • Define and evaluate performance metrics
  • Conduct hypothesis testing on experiment data
  • Create visual and interactive dashboards in Tableau
  • Present clear, data-driven insights

📦 Deliverables Checklist

  • Cleaned data files
  • [ 01_EDA_and_Cleaning.ipynb ] Python analysis notebooks
  • [ Dashboard_V1_.twb.twbx ] Tableau dashboard (.twbx)
  • [ https://prezi.com/p/edit/eqsnwx6qvkuk/ ] Final presentation slides
  • README.md (this file)

About

EDA and Data Cleaning Performance Metrics Hypothesis testing Experiment Evaluation Tableau

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors

Languages