Welcome to my portfolio showcasing a comprehensive case study for Bellabeat, using Fitbit data. This repository contains all the materials data, analysis code, visualizations, and the final presentation, to demonstrate my end-to-end analytics workflow.
Bellabeat-Project/
├── data/ # Cleaned CSVs and raw datasets
│ ├── user_summary.csv
│ ├── avg_daily_steps.csv
│ ├── avg_heart_rate_user.csv
│ ├── avg_hourly_steps.csv
│ ├── avg_mets_user.csv
│ ├── avg_user_steps.csv
│ ├── daily_steps_heart_rate.csv
│ ├── hi_minutes.csv
│ ├── high_intensity_summary.csv
│ ├── hourly_intensity_summary.csv
│ └── weight_change.csv
├── code/ # Analysis scripts
│ ├── Bellabeat_analysis.R # R script combining all summaries
│ └── ggplots_bellabeat.R # R scripts for ggplot2 charts
├── tableau/ # Tableau workbook and exports
│ └── Bellabeat_Dashboard.twbx
├── presentation/ # Final presentation
│ └── Bellabeat_Insights.pptx
├── images/ # Exported charts for README or PPT
│ ├── mets_vs_weight_change.png
│ ├── avg_act_per_min.png
│ ├── steps_vs_act_type.png
│ ├── act_vs_rest_heart_rate.png
│ └── act_int_vs_mets.png
└── README.md # This document
- METs vs Weight Change: Each +1 MET correlates with ~0.23% weight loss (R²=0.29).
- Activity Intensity vs METs: Very-active minutes explain ~60% of MET variation (p<0.0001).
- Activity vs Resting Heart Rate: No single intensity bucket strongly predicts resting heart rate.
- User Segmentation: Users grouped into High, Moderate, Low intensity for targeted recommendations.
- Highlight "Very Active" streaks in the Bellabeat app.
- Send light-activity reminders to boost overall METs.
- Create combined active-minute dashboards linking to resting HR trends.
- Introduce MET-based fitness challenges in marketing campaigns.
- Data Sources: Original Kaggle Fitbit dataset (CC0 Public Domain).
- R Code: Modular scripts in
code/folder. - Tableau: Dashboard workbook in
tableau/, View the Bellabeat Dashboard – Exercises - Presentation: Final slide deck in
presentation/.