Background Information:

Urška Sršen: Bellabeat’s co-founder and Chief Creative Officer ○ Sando Mur: Mathematician and Bellabeat’s cofounder; a 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.

1) Exploring the number of users in each dataset:

To explore the number of users in each dataset, first of all, I converted the datatype of data which is MM:DD: YYYY hh:mm: ss into MM:DD: YYYY before uploading the dataset to SQL. Then, in SQL Big Query, I ran the below codes to explore the number of users in each dataset.

You can find the link to the code below

2) Examine the relationship between total calories and total steps:

To examine the relationship between total calories and total steps, I used R for analysis. First of all, I installed below the necessary packages in R:

Result:

The result shows that total steps and calories have a positive relationship, the higher the calories user consume, the higher the total steps that users walk or run.

3) Examine which day users are most active in walking and which day users are least active in walking.

I also examined which day users consumed highest calories and which day users consumed lowest calories. To do this, I used Excel to change the activity date into weekday by using text formula. After that, using SQL, I could answer the question which weekday users had the highest average number of steps and highest average distance.

The result shows that Saturday is the day that users are most active walking, followed by Tuesday. The result also indicates that Saturday and Tuesday are also the days that users have the highest calories on average, meanwhile, on average, Sunday is typically not the day that users are active in walking.

Result:

Visualization on Tableau:

hourly intensities

To easier examine this relationship, I used SQL to merge data between TotalSteps and TotalMinutesAsleep by Id level

6) Examine what time users are most active.

I used hourly intensitiesinclude data to answer this question. First, to easily input in SQL, I opened this file in Excel and split ActivityHour data that combines both Day and Time to get only ActivityHour which was not including the date, month and year.

It can be seen that users are most active from 17-19h, which indicates that users will often exercise during this period of time.

Conclusion and recommendation:

My analysis above has provided some interesting marketing insights to Bellabeat company. Below is the summary of my findings and recommendation:

1. More users prefer to use daily Activity feature to track their wellness habit than other features such as sleep, heartrate or weight. The second most popular feature is sleep, the third popular one is heartrate, followed by weight. This information can be used in featuring Bellabeat products.

2. Saturday is the day that users are most active in walking or running, followed by Tuesday. These 2 days are also the days that users consume the highest calories in the entire week.

3. Calories and Steps have a positive linear relationship. When users run or walk more often, they will also consume more calories.

4. Sedentary Minutes and TotalSteps does not have a linear relationship. But it can be observed that number of steps will increase when sedentary minutes are from 0-500 mins, and number of steps will reduce when sedentary minutes are from 1000-1500. This is interesting, because we can recommend users who have sedentary time more than 1000 minutes should start exercising more.

5. There is, however, no relationship between Total Steps and Sleep duration. Exercising more does not help users to sleep longer time.

6. It can be seen that users are most active from 17-19h, which indicates that users will often exercise during this period of time. In Bellabeat app, we can recommend users to practice exercise during this period of time. The data can actually be adjusted to each user data and we can advise excerse time for each user.

7. It would be better if this Fitbit dataset contains more demographic information of users such as their gender. Fitbit also does not have information about user hydration level, so Bellabeat can incorporate this information into their product and provide suggestions regarding suitable hydration level to its users.

Tableau Link:

Tableau Dashboard

GitHub Link:

SQL and R Code