Bellabeat is a high-tech manufacturer of health-focused products for women. It is a successful small company, but has the potential to become a larger player in the global smart device market. Urška Sršen, cofounder and Chief Creative Officer of Bellabeat, believes that analyzing smart fitness device data could help unlock new growth opportunities for the company. In this project I've analyzed smart device usage data from non-Bellabeat products to gain insights on how people are using these smart devices and then selecting a Bellabeat product applied these insights to create high level recommendations for Bellabeat marketing strategy.

1. Data Cleaning and Processing
2. Exploratory Data Analysis
3. Descriptive Statistics
4. Data Visualization

• Urška Sršen: Bellabeat’s Cofounder and Chief Creative Officer
• Sando Mur: Mathematician and Bellabeat’s Cofounder; key member of the Bellabeat Executive Team
• Bellabeat marketing analytics team: A team of data analysts responsible for collecting, analyzing, and reporting data that helps guide Bellabeat’s marketing strategy. I joined this team six months ago and have been busy learning about Bellabeat’s mission and business goals — as well as how I, as a junior data analyst, can help Bellabeat achieve them.

• Bellabeat app: The Bellabeat app provides users with health data related to their activity, sleep, stress, menstrual cycle, and mindfulness habits. This data can help users better understand their current habits and make healthy decisions. The Bellabeat app connects to their line of smart wellness products.

• Leaf: Bellabeat’s classic wellness tracker can be worn as a bracelet, necklace, or clip. The Leaf tracker connects to the Bellabeat app to track activity, sleep, and stress.

• Time: This wellness watch combines the timeless look of a classic timepiece with smart technology to track user activity, sleep, and stress. The Time watch connects to the Bellabeat app to provide you with insights into your daily wellness.

• Spring: This is a water bottle that tracks daily water intake using smart technology to ensure that you are appropriately hydrated throughout the day. The Spring bottle connects to the Bellabeat app to track your hydration levels.

• Bellabeat membership: Bellabeat also offers a subscription-based membership program for users. Membership gives users 24/7 access to fully personalized guidance on nutrition, activity, sleep, health and beauty, and mindfulness based on their lifestyle and goals.

Analyze smart device usage data from non-Bellabeat products to gain insights on how people are using these smart devices and then selecting a Bellabeat product to apply these insights to create high level recommendations for Bellabeat marketing strategy.

Produce a report with the following deliverables:

1. A clear summary of the business task
2. A description of all data sources used
3. Documentation of any cleaning or manipulation of data
4. A summary of the analysis
5. Supporting visualizations and key findings
6. High-level content recommendations based on the analysis

For this project I'll focus on the Bellabeat app. We want to know how FitBit users use FitBit to keep a track of their health. The data provided here may help us find the answer to the following questions:

1. What are some trends in smart device usage?
2. How could these trends apply to Bellabeat customers?
3. How could these trends help influence Bellabeat marketing strategy?

FitBit Fitness Tracker Data from Kaggle is the primary source for this project. It consists of 18 files that contain various information metrics based on which we can provide recommendation to the BellaBeat stakeholders.

For this project I'll be using only 2 files and later merge them to create one file to work upon-

• dailyActivity_merged.csv
• sleepDay_merged.csv

I used the R0CCC method to check the integrity and reliability of the data.

1. Reliability:
   The data set used for this project is provided in Kaggle. It contains personal fitness tracker data from 30 fitbit users. There is no information about the age, gender, profession, accuracy or bias during data collection.

2. Originality: This data was collected by Amazon Turk between 12/03/2016 - 12/05/2016 and has been preprocessed and uploaded by Kaggle User Mobius. The data set is licensed under Creative Commons which claims no copyright of the data and allows copy, modification and performing work, even for commercial purposes, all without asking permission. Since the data was collected from third party, the originality cannot be vetted.

3. Comprehensiveness: 30 eligible FitBit users consented in submission of their health tracker data, that includes minute-level output for physical activity, heart rate, sleep monitoring. However, the dataset contains many limitations:

   • There's no information about the person's age, gender, height, weight, profession etc, that could help us identify patterns of activity of individual users. We are not sure if the sample is representative of the population as a whole.
   • There is no information about the unit of distance.
   • There are no information about intensity, its unit and how it is measured.
   • There are inconsistencies in logged data, not all 30 users have data for full 30 days.
   • The data contains 1440 minutes of sedentary minutes which is equivalent to a day, indicating that they were not wearing the tracker throughout the day.
   • The data contains information from March to May. Since the data spreads for both April and May, there are participants with 31 days of data, making the time-frame for data analysis inconsistent.

4. Citation: Furberg, R., Brinton, J., Keating, M., & Ortiz, A. (2016). Crowd-sourced Fitbit datasets 03.12.2016-05.12.2016 [Data set]. Zenodo. https://doi.org/10.5281/zenodo.53894

5. Current: This data was collected from 12/03/2016 to 12/05/2016, which indicates that the data is quite outdated. It decreases the data's usability.

1. Choosing the tools: For data processing and cleaning I used R programming language and RStudio as IDE.

2. Checking the data for errors:

   1. After importing the files, saved them in different data frames activity and hourly_sleep respectively.
   2. Checked their data types using spec() function.
   3. Checked for null values. There was none.
   4. Checked for duplicates, and removed 3 duplicate entries from hourly_sleep dataset.

3. Transforming the data to work with it effectively:

   1. Changed the column names for more consistent and cleaner names using clean_names() function from janitor package. It changed the CamelCase names to lower_snake_case names.
   2. Created two copies of the cleaned data. Used the copied data for analysis.

4. Document the cleaning process: All the data-processing and transforming steps been documented in this report as well as in the code.

The following steps are taken to analyse the data-

1. Aggregate the data: Merged the two datasets into one final data set named daily_activity_sleep.

2. Organize and format the data: Selected the only columns required for the analysis.

3. Performed descriptive statistics on the numeric columns and visualized the results through histograms.

4. Created a correlation matrix to find out the strength of relations between various variables.

5. Identified the following metrics to evaluate the performance of fitness tracking system- 1. total steps taken 2. calories burned 3. active minutes 4. sedentary minutes 5. total minutes asleep

6. Based on total daily steps taken, grouped the users into 4 categories -
   1. Sedentary- mean daily steps is less than 5000
   2. Lightly active-mean daily steps is between 5000 and 7499
   3. Fairly active- mean daily steps is between 7500 and 9999
   4. Very active - mean daily steps is greater than 10000

7. Based on total minutes asleep, I've grouped the users into into 3 categories -
   1. Bad Sleeper - mean daily sleeping time is less than 360 mins (6 hours)
   2. Normal Sleeper - mean daily sleeping time is between 360 mins (6 hours) to 480 mins (8 hours)
   3. Over Sleeper - mean daily sleeping time is greater than 480 mins (8 hours)

8. Identify trends and relationships:

To find out the strength of relations between various variables I've created a correlation matrix. A correlation matrix is a statistical technique used to evaluate the relationship between two variables in a data set. The matrix is a table in which every cell contains a correlation coefficient, where 1 is considered a strong positive relationship between variables, 0 a neutral relationship and -1 a strong negative relationship.

From the correlation matrix we can observe the followings-

1. Calories burned is positively correlated with total steps taken, total distance travelled and total activity minutes. The scatterplots shows their relationship visually.

   

   

   

2. Sedentary minutes is negatively related with total steps, total distance and lightly active minutes.

3. Active minutes have positive but weak correlation with total steps, total distance, fairly active minutes and calories burned.

4. Total minutes asleep is positively related with total time in bed.

Although the data set had lots of limitations, namely lack of reliablity, small number of users, lack of proper details etc, this short exploratory data analysis was able to identify some usage pattern of FitBit Fitness Tracker users.

• The average daily steps was 8108, close to the CDC (Center for Disease Control & Prevention) recommendation, but lower for lightly active and sedentary users.

This means that most users walked around 8,000 steps per day, which is close to the CDC's recommendation of 10,000 steps per day for adults. However, this average was dragged down by the lightly active and sedentary users, who walked less than the average. This suggests that these users need to increase their physical activity to improve their health.

• The average very active minutes was 24, with most users having 0 - 10 minutes.

This means that most users spent only 24 minutes per day doing vigorous physical activities, such as running, cycling, or playing sports. Moreover, most users had even less than 10 minutes of very active minutes per day. This is far below the CDC's recommendation of 75 minutes of vigorous physical activity per week for adults. This implies that most users need to boost their intensity of exercise to achieve the health benefits of physical activity.

• The average sedentary minutes was 799 ~ 13.31 hours, possibly due to tracker not being worn, and higher for sedentary users (~ 19 hrs) than fairly active users (~13 hrs).

This means that most users spent around 13 hours per day being inactive, such as sitting, lying down, or sleeping. This could be partly because some users did not wear their tracker all the time, so the tracker did not record their movements. However, this could also indicate that some users had a sedentary lifestyle, which is associated with increased risks of obesity, diabetes, heart disease, and other chronic conditions. The sedentary users had the highest average of sedentary minutes, around 19 hours per day, while the fairly active users had the lowest, around 13 hours per day. This shows that there is a significant difference in the amount of time spent being inactive among different types of users.

• The average calories burned was 2323, lower than the average for women and men, and lowest for sedentary users.

This means that most users burned around 2300 calories per day, which is lower than the average calorie expenditure for women (2000 calories) and men (2500 calories) in India. This could be because most users did not engage in enough physical activity to burn more calories. The sedentary users had the lowest average of calories burned, around 1900 calories per day, which is not enough to maintain a healthy weight. This suggests that these users need to increase their calorie intake and/or their physical activity to balance their energy expenditure.

• The average minutes of sleep was 419 ~ 7 hours, more for lightly active users than sedentary and very active users.

This means that most users slept around 7 hours per night, which is close to the recommended amount of sleep for adults (7 to 9 hours) by the National Sleep Foundation. However, this average was influenced by the lightly active users, who slept more than the average, around 8 hours per night. The sedentary and very active users slept less than the average, around 6 hours per night. This indicates that these users may not get enough sleep to support their health and well-being. Sleep is essential for the proper functioning of the immune system, brain, heart, and other organs. Lack of sleep can impair cognitive performance, mood, memory, and alertness. Therefore, these users may benefit from improving their sleep quality and quantity.

From the FitBit Fitness Tracker data we can infer that the users primarily uses these devices to track their daily calorie burn, physical activity data, heart rate, hydration level and sleeping pattern. The Bellabeat app can consider the following matrices to provide accurate information about the daily activities of the users:

• Daily total steps
• Daily calories burned
• Daily active minutes
• Daily sedentary minutes
• Duration of sleeping (light, normal or over sleep)

As these variables are closely related to each other, one factor affects the others. Using the information from those matrices the Bellabeat app can further give information about user's weight, heart rate, cholesterol level, hydration level and suggest healthy eating habits, recommend physical exercises and back up their claims by providing scientific facts.

• As users do more physical activity, they burn more calories, which in turn, helps them lose more weight. In addition to giving daily updates on calorie burning, Bellabeat can suggest some ideas for low-calorie intakes.

• If users want to improve their sleep, Bellabeat should consider using app notifications to go to bed and notify them if they've had enough sleep or if they overslept.

• The Bellabeat app can notify the users when they're sitting (sedentary time) for more than that is healthy, and remind them to do some light activities like, taking a stroll outside, or stretching up a bit etc.

• When a user does the recommended activities, the Bellabeat app can show motivational messages and reward them. That way the users will be encouraged and they'll continue to do the hard work.

• Social media marketing : Bellabeat can use social media platforms like Facebook, Instagram, and Twitter to reach potential customers by creating engaging posts that highlight the benefits of their products and services and run targeted ads that reach people interested in fitness and wellness.

• Influencer marketing : They can collaborate with fitness influencers to promote their products and ask them to review them on their social media channels.

• Email marketing : Bellabeat can use email marketing to keep their customers engaged and informed by sending regular newsletters that provide tips on fitness, wellness, and healthy eating and personalized emails that recommend products based on the user's activity level, sleep patterns, and other health data.

• Gamification : They can use gamification to make their app more engaging and fun by creating challenges that encourage users to be more active, sleep better, and eat healthier and offer rewards like badges, trophies, and discounts to users who complete these challenges.

• The time of the day was not considered, so the activity patterns were not clear.
• The data set was small, outdated, and unreliable, so the findings were not generalizable.
• The relation between steps, calories, and sleep was not further investigated, to avoid bias and analysis paralysis.
• There was no user data for Bellabeat products, for more direct and effective analysis.

• spec()
• Janitor Package
• merge()
• n_distinct()
• Correlation Matrices
• Correlation Data Frame
• Histogram in r
• Higher Daily Step Count Linked with Lower All-cause Mortality
• How much sleep do I need?