What is Feature Engineering?
When we look at numbers and metrics like the ones that show up on the watch what am I supposed to do with it?
This week after a gap of 15 days:
-
I ran 5 kilometers.
-
Couple of weight training sessions that burnt 300 calories each
-
Over the weekend I had a night out with friends and where there where quite a few drinks consumed leading to horrible sleep that night.
I have all of these number stats showing up How am I doing ? how fatigued am I, and how well-trained is my body ?
Feature engineering is the process of transforming raw health data into meaningful and actionable insights. It bridges the gap between raw metrics and analytical models by uncovering patterns, enhancing data interpretability, and improving decision-making.
By converting raw data into key features that highlight trends and correlations, feature engineering makes data easier to grasp, enabling better data-driven health and fitness decisions.
How Does Feature Engineering Enhance Garmin Data?
Feature engineering refines raw metrics into actionable insights through techniques such as:
-
Temporal Analysis: Identifying trends over weeks or months to uncover long-term health patterns.
-
Personalized Metrics: Creating tailored insights like fitness-to-fatigue ratios or training load assessments.
-
Clustering: Grouping similar data points, such as categorizing activity types or sleep patterns.
-
Simplification: Using methods like PCA to focus on key patterns in complex datasets.
By applying these techniques, raw Garmin data becomes a powerful tool for understanding health trends
TL;DR: Activities to Training:
Turning Daily Activities into Traning Insights
Our everyday activities—like Running, Cycling, Strength Training, and Yoga—paired with metrics such as Steps, Calories, and Heart Rate, are transformed into:
Key Training Metrics:
-
Training Load: Measures the overall impact of your activities.
-
Monotony: Tracks the variability in your training routine.
-
Strain: Captures accumulated stress over time.
These metrics create a strong foundation for assessing Performance
TL;DR: Performance to Health
Use Training Metrics like Training Load, Monotony, and Strain to unlock performance metrics such as:
-
Training Load: Quantifies overall effort.
-
Fitness: Reflects long-term training progress.
-
Fatigue: Indicates short-term exertion.
-
Stress: Measures physiological impact from activities.
These form the foundation for Health Metrics such as :
-
Strain: Tracks cumulative stress to prevent overtraining.
-
Fitness: Assesses overall physical conditioning.
-
Seasonality: Highlights quarterly trends for goal alignment.
-
Base Fitness: Offers insights into long-term endurance.
Let’s make these metrics now work for us for a more holistic view of our fitness:
interactive graphs on desktop version
TL;DR: Sleep into Detials:
Turning Sleep into Sleep Insights
-
Sleep Score: Quantifies overall sleep quality by assessing duration, depth, and efficiency.
-
REM Sleep: Crucial for cognitive recovery, emotional balance, and memory consolidation.
-
Deep Sleep: Facilitates physical restoration and muscle repair.
-
Awake Time: Reflects sleep interruptions, affecting overall sleep efficiency.
These metrics create a strong foundation for assessing Performance
TL;DR: Sleep into Sleep Quality
Sleep Sentiment:
Negative sentiment aligns with restless or poor-quality sleep, often associated with stress or insufficient recovery.
Positive sentiment corresponds to refreshing or deep sleep, indicating optimal rest and recovery.
Sleep Duration:
Short durations are linked with poor structure and insufficient REM, which can negatively impact recovery.
Longer durations tend to correlate with optimal structure, deep sleep, and continuous patterns, contributing to better overall rest.
Sleep Quality:
Factors like poor quality, restless sleep, or not enough REM can highlight disruptions, which are often tied to stress or misaligned routines.
High-quality sleep categories like recovering, refreshing, or continuous reflect balanced routines and effective recovery.
Daily Context:
Stress and lack of evening rest contribute to poor-quality sleep, emphasizing the importance of managing external factors.
Positive contexts like exercise and good timing enhance sleep quality by promoting restorative patterns.
HRV Nightly Avg: Monitors nightly HRV trends.
HRV Weekly Avg: Captures broader HRV trends across the week.
HRV 5min High: Highlights peak HRV moments, valuable for identifying optimal recovery phases.
HRV Baseline Range: Establishes a personal baseline to gauge fluctuations in recovery and stress.
HRV Range: Displays the variability between high and low HRV metrics, aiding in spotting anomalies.
Average Heart Rate: Tracks your overall heart rate during the sleep cycle, reflecting recovery and stress levels.