To approximate this data, this case study is an analysis of a public datset of 33 individuals whose activity was recorded by their Fitbit devices over a period of two months. This data provides users' calories burned, activity intensity, step count, and sleep behavior, aggregated over days, hours and minutes. This dataset can be found here.

To better focus this analysis we will look for trends in user data across fitness level, general useage, and step count aggregated daily. We will also look for activity trends throughout the day, such as whether users are more active in the morning or evening.

Additionally, some of the provided data, such as the weight log files, are unusable for analysis due to user inconsistency. Out of 33 users, only 2 were consistent in manually adding their daily weight to the app throughout the 30 day sampling period, so unfortunately this analysis will be unable to provide insights on how fitness tracker wearers may be using their devices for weight loss.

Data was inspected, organized, and cleaned in Microsoft Excel before being loaded into Tableau for visualization.

This dataset has been visualized as an interactive dashboard in Tableau which can be accessed here.

Initial analysis of user activity shows that users are primarily sedentary when wearing fitness devices. The average user is sedentary for 81% of the day, lightly active for 16% of the day, anf fairly or very active for 1% and 2% of the day, respectively. This observation makes logical sense, since sleep takes up one third of a 24-hour period and users are likely wearing their devices all the time. Many users may additionally have sedentary office or administrative jobs, which could account for the remainder of inactive time.

Eliminating sedentary time from the analysis gives a better picture of what type of activity users prefer when they choose to work out: light activity, moderate activity, or very active (rigorous) activity. Users clearly have a preference for light activity like walking, yoga, or pilates. These exercises are almost universally beneficial, low-impact, and easy to incorporate into one's daily routine. These are good options for people seeking to build a more active lifestyle, especially for those exploring how to live more healthfully for the first time. Interestingly, more rigorous workouts (like running, kickboxing, or HIIT) are the next most popular workout type and make up nearly 10% of user activity outside of sedentary behavior.

Analysis of average daily step count over a 24 hour period serves as a measure for what time of day users prefer to be active. The two most popular times of day for users to be active are in the middle of the day (perhaps on lunch break from work) and immediately after the workday, with the highest average user activity recorded during the 6:00pm hour. Users were rarely active during the late night and very early morning hours when people are usually asleep. Overall activity records reflects a diurnal circadian rhythm, with users moving around during the daylight hours generally, and peak activity points during the lunch hour (noon to 2pm) and immediately after work (5pm to 7pm).

Regarding the range of average step counts across the entire user pool: the user with the highest average step count took 16,040 steps per day and the user with the lowest average step count took only 916 steps per day. The median user took 7,283 steps per day on average. The most popular day for users to be active was Saturday, with an average of 344 steps across the user pool. The least popular day for activity was Sunday, with 288 steps on average among users.

Step counts were aggregated by using the recommendations found here to classify users according to their activity levels. Users overall tended to favor the less active portion of the spectrum of activity, with 24% of users considered sedentary (<5,000 average daily steps), 27% lightly active (5,000 to 7,499 steps), 27% somewhat active (7,500 to 9,999 steps), 15% active (>10,000 steps) and 6% very active (>12,500 steps).

The above trends in how consumers use wearable fitness devices (FitBit) provides substantial insight into how Bellabeat customers might use their devices. Based on these findings, some recommendations for the Bellabeat marketing team are:

1. Provide some social aspect of the app so that users can connect (or compete!) with friends to improve step counts and achieve fitness goals.
2. Issue discount codes for users to invite their friends to buy Bellabeat products and get fit together.
3. Keep track of users' overall activity levels and celebrate when a user moves into a higher category (ie from sedentary to lightly active).
4. Push step count reminders to users on days and times when they are already likely to want to be active: Saturdays and weekdays after work.
5. Reward users who consistently meet daily and weekly step count goals.
6. For users who are interested in weight loss, perhaps the introduction of a digital scale that pairs with the Bellabeat device via Bluetooth would help users stay consistent about tracking their weight and remove the barrier of needing to add this data to the device manually.