BellaBeat Capstone

Say Hello to Bellabeat!

What is BellaBeat?

Bellabeat is a sucsessfull company that produces technology aimed at improving womens health and fitness. Bellabeat has a variety of different product such as a wrist watch, water bottle, jewlery, as well as the app and the membership. Each product aims to improve a different aspect of female health such as water intake, cycle tracking, sleep, stress, and overall activity levels.

My Role

As a data analyst at BellaBeat I have been tasked with analyzing and comparing data from different smart devices to help improve Bellabeats sales. To do this I will be looking at fitbit tracking data from Kaggle and have decided to specifically analyze the data that pertains to Bellabeats leaf product. The leaf product can be worn as necklace, bracelet, or clip and tracks stress, sleep, and activity levels of the user.

Systems Used

I used google sheets to make my analysis. The data sets that I decided to use are dailyActivity_merged.csv, heartrate_seconds_merged.csv, and minuteSleep_merged.csv. These datasets all pertain to the trafficking purposes of the leaf.

What did I find and how did I clean the data?

Heart Rate Data

The heart rate data was recorded by various increments of seconds throughout the day. There were thirteen different fit bit users that the data was logged from. The average heartrate throughout the thirteen users and throughout the day was 86 BPM. This is considered on the higher half as the averge range of BPM is 60-100. The range of BPM within the dataset is 36 to 185. Heart rate can pertian significantly to stress rates so we can come to understand that higher heart rates can correspond to higher levels of stress. It is hard to determine the route cause of the spike in heartrate if it was activity or if it was stress, but in general the higher average heart rates could correlate to higher stress.

Sleep Data

The sleep data was very difficult to try and clean as hours asleep were listed in a fairly complicated fashion. Each seperate hour was logged alongside the date and id of the fitbit user. I adjusted the data set to be able to count the hours spent sleeping and from what hours from 37 different recorded dates. This is not all of the data points but a substatial amount. With a dataset like this it would be useful to utilize R and use the seperate function to seperate these values. From what I viewed in the dataset peoples sleep schedules were either inconsistent or they were consistent yet not recieving enough sleep.

Daily Activity

When skimming the activity levels for fit bit users the daily calories burnt average to 1378.5 for users surveyed. If the 1378.5 calories were all active calories then that would indicate high levels of acticity however the dataset is compiled of a variety of different types of activity such as very, fairly, or lightly active excercise. The average of very active minutes is 16.5 minutes which is fairly average as you are expected to meet 75 minutes of rigorous activity a week. Fairly active minutes average at six minutes which is below the suggested thirty minutes of excercise daily. Lightly active minutes average roughly 131.5 minutes per day. However, sedentary minutes have the highest average at about 667.5 minutes. This statistic isn’t entirely unexpected as many of us live sedentary lifestyles due to our jobs and general societal norms. Additionally, time spent sleeping is logged within sedentary minutes.

My Suggestions for Stakeholders

My best suggestions for stakeholders would be to focus on activity levels as well as sleep hours. Both of these variables are generally very easy for users to comprehend and understand. Stress levels unless presented with a corresponding rate could be difficult for users to understand. Additionally users may wonder how the device determines stress levels to begin with. Another suggestion is that I think it is important to convey to users active calories burnt so they know what activity threshold they are performing at using sedentary calories burnt makes it difficult to see if users are really challenging themselves during activity or a workout. Lastly, advocating for proper sleep is a huge proponent of the BellaBeat company. This coupled with the fact that many people have inconsistent sleep schedules (at least those within the dataset) leaves a great marketing opportunity for BellaBeat to market and improve sleep worldwide.