1 of 19

CASE STUDY

DEMONSTRATING SKILLS USING TABLEAU

CHRIS UMBANHOWAR

bellabeat

2 of 19

AGENDA

bellabeat

3 of 19

BACKGROUND

Bellabeat, a high-tech manufacturer of health-focused products for women

​

Their products vary from:

    • The Bellabeat app - providers uses with health data such as their activity, sleep, stress, menstrual cycle, and mindfulness habits.
    • Leaf tracker & Time watch - both products connect to the Bellabeat app to track activity, sleep, and stress.
    • Spring water bottle - used to track your daily on water and connects to the Bellabeat app to track hydration levels
    • Membership - a subscription-based membership used to give fully personalized guidance on nutrition, activity, sleep, health and beauty, etc.

​

4 of 19

OBJECTIVE

Spring is Bellabeat’s smart water bottle that nudges users to stay hydrated throughout their day. To position Spring most effectively, we’ll lean on patterns in how Fitbit users engage with their trackers—when they’re most active, when they rest, and how frequently they interact with their devices—to tailor hydration prompts and messaging.

​

Business Task:

​

Analyze hour-level Fitbit activity, heart-rate, and sleep data to uncover daily engagement rhythms, then translate those insights into a data-driven marketing strategy for Spring.

5 of 19

OBJECTIVE

Key Questions:

​

    • When during the day are users most active or sedentary—i.e., when would hydration reminders land best?
    • How do sleep and heart-rate patterns suggest times of peak dehydration risk?
    • Which user segments (e.g., highly active vs. lightly active days) should Spring target with customized messaging?

​

Desired Outcome:

​

A set of high-level recommendations on:

    • Timing of Spring’s in-app or push notifications
    • Messaging themes that resonate (e.g., post-workout boost, mid-afternoon slump)
    • Channel prioritization (email, app alert, social ads) to maximize bottle adoption and ongoing engagement

6 of 19

Participants: 30 Mechanical-Turk Fitbit users, March 12–May 12, 2016 (Observations counted in dailyActivity_merged)

​

Data Types: Minute-level and aggregated outputs for:

    • Physical activity (steps, calories)
    • Heart rate (bpm)
    • Sleep state (minute-by-minute asleep/awake)

​

Objective: Understand daily engagement rhythms to inform Spring’s hydration prompts

​

Kaggle Data Source: FitBit Fitness Tracker Data

DATA OVERVIEW

7 of 19

    • Two survey waves (3/12–4/11 and 4/12–5/12), each unpacked into 11 CSVs
    • Key files used:
    • dailyActivity_merged.csv – daily totals (steps, active/sedentary minutes in a day, calories)
    • hourlySteps_merged.csv – step counts by hour
    • hourlyCalories_merged.csv – calories burned by hour
    • heartrate_hourly_merged.csv – average heart rate by hour (aggregated from seconds data)
    • minutesSleptPerHour_merged.csv – minutes asleep by hour (aggregated from sleep flags)

​

3. Final dataset: Five sheets in bellabeat_dataset.xlsx, covering every hour for a two-month window (dailyActivity_merged.csv will not cover every hour)

​

DATA SOURCES & STRUCTURE

8 of 19

    • Timestamp standardization: Converted all time fields to proper Date/Time types
    • Null & duplicate removal: Dropped blank rows and de-duplicated on (Id + timestamp)
    • Aggregation to hourly grain:
    • Heart-rate seconds → averaged per minute → then per hour
    • Sleep flags → summed minutes asleep per hour

4. Derive flags:

    • HighExertionFlag on hourly calories (Calories > 150)
    • WakeUpFlag on minutes-asleep transitions

​

​

DATA CLEANING & PREPARATION

9 of 19

Append two periods (3/12–4/11 + 4/12–5/12) via Power Query → one sheet per metric:

​

    • dailyActivity_merged
    • hourlySteps_merged
    • hourlyCalories_merged
    • heartrate_hourly_merged
    • minutesSleptPerHour_merged

​

​

​

​

DATA MERGING

10 of 19

    • Merged all hourly tables on Id + Time to align activity, heart rate, and sleep in Tableau

​

    • Visualized in Tableau Public to uncover temporal patterns and segment behaviors

​

​

​

ANALYSIS

View visualizations from Tableau Public Profile -> Bellabeat Spring Analysis

11 of 19

DATA SUMMARY

    • When during the day are users most active or sedentary—i.e., when would hydration reminders land best?

​

    • How do sleep and heart-rate patterns suggest times of peak dehydration risk?

​

    • Which user segments (e.g., highly active vs. lightly active days) should Spring target with customized messaging?

​

12 of 19

KEY FINDINGS

13 of 19

14 of 19

15 of 19

16 of 19

17 of 19

18 of 19

1. Timing of Notifications

    • Morning Wake-Up (5–9 AM): Send the first reminder as users transition from sleep—capitalizing on Heart Rate spike and low sleep overlap. (Slides 14 & 15)
    • Midday Slump (1–3 PM): Gentle nudge during activity lull to break sedentary behavior. (Slides 12 & 13)
    • Evening Cool-Down (7–9 PM): Post-activity prompt when users finish their second daily step peak. (Slides 12 & 17)

2. Messaging Themes

    • Post-Workout Boost: “You crushed that workout—refuel with a refreshing sip!” (targets high-activity users)
    • Afternoon Reset: “Feeling the slump? Some oz. from the Spring can recharge your afternoon.” (broad audience)
    • Wind-Down Wellness: “Hydrate now to support overnight recovery.” (evening reminder)

3. Channel Prioritization

    • App Push Notifications: For time-sensitive, in-moment nudges (Morning & Evening windows).
    • Email Digests: Daily summary with personalized hydration stats & tips (Morning recap).
    • Social Media Ads: Targeted Facebook/Instagram stories around peak hours for broader brand awareness.

​

RECOMMENDATIONS

19 of 19

THANK YOU

www.chrisumbanhowar.com

https://www.linkedin.com/in/chris-umbanhowar-60a359209/