Data analytics and metrics optimization in wearable electronics
Overview
In wearable devices, the delivery of metrics and insight from the usage of those devices is the most important aspect of user experience and satisfaction. Users expect to get certain insights and recommendations from their wearable electronics that will support their daily activities and they need to be accurate. All of that is achieved by proper utilization of user-generated data that tracks those users’ habits and behavior.
For Bellabeat, the producer of smart jewelry for women, we came on as collaborators to analyze their data, optimize and discover new metrics presented to users. The main goal was to uplift user experience and satisfaction. Our approach was to explore their user-generated data and see which influence metrics the most so we can enhance accuracy and relevancy.
We wanted users, in this case women, to get certain numbers and insights that follow their lifestyle and could support their daily activities.
Project scope
Discovery:
- Workshops
- Client interviews
- Databases inspection
- Analysis of currently implemented algorithms
Procedure:
- Data cleansing algorithms implementation
- Data analysis
- Implementation of algorithms for sensory data interpretation
Methods:
- Data cleaning (detecting, and fixing or removing noisy data)
- Statistical data analysis for each user that enables individualized metrics calculation
- Design and improvement of algorithms for calculating user activity and sleep time
- Implementing methods from scientific papers
Project duration:
- 9 months
Technologies used:
Python
Pandas
NumPy
Scikit-learn
Jupyter Notebook
Main goals:
Main issues:
Main goals:
- Recognize and implement new metrics
- Optimize metrics that evolve the same way users evolve
- Improve user experience and satisfaction
- Improve data accuracy, completeness, and interpretation
Main issues:
- Lower frequency of data sampling
- Noise in data
- Inaccurate and missing data
- Differences and peculiarities in device usage by users with different lifestyles and habits
Our approaches:
In any wearable electronics, the main focus is on data that it collects from its users. Those are daily intakes of information on the user’s heart rate, breathing, activity, sleep, and others. But without proper data collection and analysis, those devices can’t deliver accurate metrics and insights that serve as a guide in users’ daily activities. Those metrics drive user satisfaction and experience and they determine if the user will keep wearing the device or not. To accomplish that, data has to be properly cleaned, complete, and interpreted the right way.
Solution:
For Bellabeat, the producer of smart jewelry, they’ve enlisted us as cooperators to achieve bigger effectiveness and accuracy of metrics from their devices that are designed to support women in their daily lives. Our approach was to get into data and find new possibilities to utilize it to deliver bigger and more impactful insights to users. The tasks consisted of defining more precise metrics that took certain outliers and errors in data into account. Another task was to discover new actionable metrics that could be presented to users in their app. The main objectives were to improve metrics, engage users, and improve user experience.
Results:
Elimination of noisy data
Discovery of new relevant data
Improvement of existing metrics
Discovery and implementation of new metrics
Bigger user engagement
Amplified user experience
Admired by
