All 11 numeric measures have complete data (8,500 non-null values each). The summary table from the profiling query is rendered above. Here is how to read it. **Skewed measures (mean well above median)** - **caffeine_post_5pm_mg** is the most skewed. The median is 0 mg but the mean is 33.2 mg (SD 51.9, range 0–250). Most people have no late caffeine, and a minority drive the average up. Its spread is about 1.6 times its mean, the widest relative spread of any column. - **next_day_fatigue_score** has a mean of 3.79 against a median of 3.2 (range 1–10, SD 2.69). Most scores are low, with a tail of high-fatigue people. - **bedtime_phone_minutes** has a mean of 59.3 against a median of 51 (range 1–180, SD 36.9). It is right-skewed, and the maximum of 180 looks like a cap. - **sleep_latency_min** is mildly right-skewed, with a mean of 40.7 against a median of 37.6 (range 6–123.3, SD 17.4). - **age** is mildly right-skewed. The mean is 34.4 against a median of 32, over a range of 18–65 (SD 11.6), so the sample leans young. **Roughly symmetric measures (mean ≈ median)** - **total_sleep_hours:** mean 6.27, median 6.33, range 3.2–9.8, SD 1.45. - **deep_sleep_pct:** mean 21.7, median 21.9, range 8.1–28.0, SD 3.07. - **rem_sleep_pct:** mean 19.1, median 19.1, range 9.6–27.0, SD 2.55. - **screen_brightness_pct:** mean 55.2, median 55, range 10–100, SD 19.9. - **physical_activity_min:** mean 35.7, median 35, range 0–112, SD 20.8. - **morning_alarm_snoozes:** mean 2.77, median 3, range 0–7, SD 1.70. **Relative spread** Dividing SD by mean (the coefficient of variation) puts the columns on one scale: - **Most variable:** caffeine (about 1.56), then fatigue (about 0.71), then phone minutes (about 0.62) and alarm snoozes (about 0.61). - **Least variable:** deep sleep % (about 0.14) and REM % (about 0.13), followed by total sleep hours (about 0.23). Sleep-stage percentages are tightly clustered. Behaviours like caffeine and phone use vary far more across people. `blue_light_filter_active` is stored as a number but is a 0/1 flag, so I treated it as a category and left it out of this profile.