Bellabeat Case Study

Google Data Analytics Certificate — Capstone Project

Author: Peter Christopher Chester | Date: November 2025

---

Overview

This case study explores how Bellabeat — a wellness technology company focused on women's health — can use smart device usage data to inform its marketing and product strategy.

Using Fitbit user data as a proxy for smart wearable behaviour, the analysis identifies a critical gap in Bellabeat's product offering and builds a data-driven argument for a new Stress/Recovery Analysis feature.

---

Business Task

Analyse competitor device trends to guide Bellabeat's marketing and product development, using Fitbit usage data and a comparative assessment of device features.

Key Question: What user behaviours in the Fitbit dataset justify a new product feature, and how should Bellabeat position it?

---

Tools Used

Phase	Tool
Data organisation & initial cleaning	Google Sheets
Deduplication, parsing, aggregation	SQL (BigQuery)
Feature engineering & merging	R / tidyverse
Visualisation	Tableau Public

---

Repository Structure

bellabeat-case-study/
│
├── README.md
├── Bellabeat_Case_Study_v.2.pdf    ← Full written analysis
├── data/
│   ├── dailyActivity_engineered - dailyActivity.csv
│   └── master_analysis_data_for_tableau (1).csv
└── scripts/
    ├── bellabeat_r_analysis.R      ← Full R workflow (cleaning → export)
    └── bellabeat_queries.sql       ← Full SQL workflow (cleaning → export)

---

Data Source

Source data: FitBit Fitness Tracker Data on Kaggle — Kaggle (arashnic/fitbit)

• 30 Fitbit users, ~3 months (March–May 2016)
• Collected via Amazon Mechanical Turk
• Public domain — free to use, modify, and distribute

Datasets used:

• dailyActivity_merged.csv
• sleepDay_merged.csv
• heartrate_seconds_merged.csv

Access the dataset directly on Kaggle: https://www.kaggle.com/datasets/arashnic/fitbit/

Key limitations: Small sample (n=30), no demographic data, collected in 2016. Findings indicate broad trends rather than statistically definitive conclusions.

---

Key Findings

1. Competitive Gap Identified

A feature comparison of Bellabeat, Fitbit, and Garmin revealed two significant gaps: Stress/Recovery Analysis and GPS. The rest of the analysis focuses on validating Stress/Recovery as the higher-priority recommendation.

2. The Compensation Discovery

The data contradicted the expected hypothesis. Users logged ~500–1,000 more steps the day after poor sleep — not fewer. This indicates widespread compensatory over-exertion: users are ignoring physiological signals for rest.

3. High Physiological Strain Segment Validated

A measurable high-risk user segment was identified — users simultaneously experiencing elevated resting heart rate and poor sleep recovery — who continued to maintain high activity levels. This group is the clearest target for intervention.

4. Feature Pivot

Based on the above, the recommendation pivots the Stress/Recovery feature from a performance optimisation tool to a Burnout Prevention and Health Alert System — better aligned with Bellabeat's wellness-first brand and its majority Moderate/Sedentary user base.

---

Visualisations

All visualisations are published on Tableau Public.

• Tableau Public profile: https://public.tableau.com/app/profile/chris.chester/vizzes
• Viz 1: Better Sleep Doesn't Change Next-Day Activity — Sleep quality vs. next-day steps comparison
• Viz 2: High Strain Validation: Users Push Harder When They Should Rest — Dual dashboard showing next-day activity and resting heart rate by strain segment

---

Recommendations Summary

1. Build Stress/Recovery Analysis as a Burnout Prevention feature — not an athletic metric
2. Target Moderate and Sedentary users (66%+ of logged days) with non-exercise interventions: guided meditation, hydration prompts, rest day suggestions
3. Align push notification timing with weekly activity patterns — motivation prompts on peak days (Mon–Sat), recovery prompts midweek and Sunday
4. Integrate phone-connected GPS for walking/hiking route tracking tied to stress and readiness scores — not pace racing
5. Position Bellabeat as the leader in whole-person wellness: "We don't just track your wellbeing — we actively help you achieve it"

---

About This Project

This is my capstone project for the Google Data Analytics Certificate. It was my first end-to-end data analysis project, taking raw Kaggle data through cleaning, SQL aggregation, R feature engineering, and Tableau visualisation — all in service of a real business recommendation.

Feedback welcome — feel free to open an issue or connect on my LinkedIn profile.