My Data Diary

Goals

1. Automatically upload data from various sources through APIs (RescueTime, Fitbit, aTimeLogger etc)
2. Visualize data in an meaningful manner
3. Change behavior
4. Maintain a timecapsule of my life
5. Predict future behavior

Motivation

I used to actively journal when I was a kid. Those entries are a timecapsule of my life; those memories would have mostly been forgotten had I no record of them. That is the motivation for this project, a more automated and perhaps more objective approach to journaling - for me and for anyone else who wants to use my templates.

I often believe I spent x time on something or didn't do y, but my data can tell me otherwise. It's through the active process of revisiting my data and idenitifying anomolies and finding interesting patterns, that I am able to tell a more cohesive story about different periods of my life and perhaps even make some changes in behavior. It's also helped me realize how finite time truly is.

Overview

• API connection to 3 data streams: Fitbit (Movement and Sleep), RescueTime (Device and Applications) and aTimeLogger (Easy data logging)
• Data aggregation
• Helper functions for visualizing data in Python
• Adaptation of the stacked bar plot for visualizing daily minutes on different activities
• Lots of charts!

Related Content

My Posts

• June 24, 2020: OpenHumans - Python Code - RescueTime Integrated Notebook
• June 16, 2020: Quantified Self - Show&Tell Slides (Recording coming soon) - Impact Covid-19 sheltering-in-place had on me
• May 16, 2020: My Site - Blog Post - How has my life changed since shelter-in-place (in charts)?
• Feb 15, 2020: OpenHumans - Self Research Post - Where did my time go? Expectations vs Reality
• Feb 14, 2020: OpenHumans - Self Research Post - How do you learn X?
• June 10, 2017: Product School/San Jose University - Invited Speaker - Presentation on My Data Dashboards in Tableau

Mentions

• June 29, 2020: Makery - External Article Mention - Open Source Body: Small data, self-research, open humans
• June 15, 2020: Tzovar.as - External Blog Mention - The effects the COVID-19 lockdown had on me
  
  Thank you to those who took a look at my project and found it interesting enough to mention.

Sample Visualizations

A two week chart of my sleep, device and active time, by the minute.

My screen time before and during the shelter-in-place (total 4 months worth of data).

Differences between the two time periods.

Running a Welch's t-test results in a pvalue of .001, meaning I am spent a lot more time on a device during shelter-in-place.

Next Steps

• Add a more detailed section in the Readme about the different files
• Incorporate heartrate data
• Extract GPS data from Google Maps
• Clean up helper function files
• Append mobile app usage with search history (RescueTime no longer captures website data)
• Store data more permanently in a database (postgres)
• Move more of data prep steps into a separate .py file

If anyone has suggestions for the above steps, especially the GPS one, I would greatly appreciate your input!

Acknowledgements

I wanted to speficially thank those who had documented/shared with me their API syntax for extracting data:

• RescueTime Fitbit API: markwk/qs_ledger
• Fitbit API: markwk/qs_ledger & amandasolis/Fitbit
• aTimeLogger API: YujiShen/TimeReport

I also wanted to give a special shout out to Mad and Bastian for creating the OpenHumans platform, a rich community for self research which served as continued inspiration to keep chipping away at this passion project of mine.