Lab Two

I’ve restructured my Apple health data to track my step counts from the years 2016 to 2020. The original data exported from the Apple health app was messy and unstructured. Using OpenRefine, I was able to clean the data to visualize the total steps for each year, month, and day of the week and understand my current health status. In addition, visualizing my daily steps for each day of the year and comparing the graphs will show my walking trends and become a way to document my peak active days. Such self-tracking data can help me improve and be more mindful of my step counts and goals for 2020.

Cleaning messy data in OpenRefine

All of the visualizations wouldn't have been possible without cleaning up my data first, as more than 161,802 records were exported with some duplicate data collected from my iPhone and Apple watch. Some of the steps I took to clean my data include:

1. Reducing 51 columns into 9, and 161,802 rows into 62,707. Renaming column headers, and deleting columns so that only the numerical value of step count is shown.
   • The mass edit feature on OpenRefine allows me to quickly edit and reduce my data, which would have been time-consuming to do without.
   • Deleting columns also makes the large dataset easier to analyze. For instance, values such as heart rate, active energy burned, distance swimming or running, and body mass can be excluded to focus on step count.
2. Getting rid of duplicate data gathered from multiple sources.
   • The text facet feature on OpenRefine allows me to see where my data is coming from (old iPhones, Apple watch, health tracker app), and exclude them to make sure they are coming from one source. If I didn’t clean it up, I would have overlapping data which will increase my total step count and produce an inaccurate result.
3. Splitting the date, time, and time zone into multiple columns.
   • Splitting the date and time allows me to isolate the data and visualize the various time references such as year, month, and date so I can aggregate them into Tableau Public.

By following these steps, I am able to open my newly clean, structured data onto Tableau Public and build a personal data dashboard to understand my walking patterns and make better habits.

Creating a personal data dashboard with my structured data in Tableau Public

2019 has been the most active year compared to the three years before. My total steps was at 2600K. Even though my steps for 2018 is a little lower than 2017, I would say that overall, I have progressed and am happy to see that I have walked more than 200,000 steps in 2019 compared to 2016. This makes me hopeful for 2020.

I seem to be the most active during the holidays in August and January. Within the past four years, I tend to hike during the summer or winter when I have a school break. September was the month with the least number of steps (around 700K), as the new semester begins, and I spend more time at home settling in or in class.

Visualizing the day of the week allows me to understand my walking trends. As expected, I tend to walk the most during the weekend. However, I am surprised to see Monday has a less total step count compared to the other weekdays, as I would expect them to be equal.

Finally, comparing my daily step counts by year will let me see my patterns and peak walking days in detail. Within the past four years, I had a total of 9 peaks that were over 25K steps a day. The most I have ever walked was 31,778 steps when I was at Disney World. Placing my data into graphs also becomes a way to document my past and remember my most active trips throughout the years. For instance, the reason why 2019 had a greater number of total steps and produced more peaks is because I was on a 4 month backpacking trip in China.

From my trend line, I can also see that I have been falling short from the recommended 10k steps a day. My 2020 step count also isn't off at a good start, with the average of 5k steps. Seeing my previous step counts compared to 2020 motivates me to raise my step counts and have more peaks than lows.

Why pursue a quantified self?

There are a lot of reasons that motivate people to track their lives. First, personal data is easy to collect. Before the project, I never realized how easy it is to attain my personal data, and how many apps were already collecting this information without my knowledge. I am thankful to have found my health data so I can document my health and behaviors for the past four years just from simply carrying my iPhone in my pocket. Even though it might not be the most accurate way to keep track of my steps, but it is by far the easiest. Second, the greatest benefit of visualizing personal health data is that it can be used for self-improvement. Seeing how my trend lines fall under the average steps of 10K, and how I have had 9 days with over 25K steps motivates me to make lifestyle changes for this year.

Here are my new goals:

1. Make sure that even on my lazy days, I will be mindful and get at least 500 steps in.
2. Be fairly consistent and try to raise my trend line closer to 10K.