**Most important metric: `next_day_fatigue_score`. Most important category: `primary_bedtime_app`.** This is my judgement, not a computed ranking. Fatigue is the outcome the dataset's decisions revolve around, and sleep latency, sleep hours and sleep-debt category all feed into it. The bedtime app is the segment a sleep-hygiene intervention could target. The figures below come from the per-app aggregates already retrieved. **How fatigue varies by app (average score, 1–10 scale):** | App | Avg fatigue score | Users | |---|---|---| | TikTok / Reels | 4.16 | 2,198 | | Instagram / Reddit | 3.94 | 1,630 | | YouTube | 3.72 | 1,928 | | Messaging / Chat | 3.59 | 847 | | Streaming (Netflix/Hulu) | 3.45 | 1,301 | | News / Reading | 3.33 | 596 | - **Spread:** TikTok / Reels users score about 0.84 points higher on average than News / Reading users (4.16 vs 3.33). The medians show the same ordering, 3.7 against 2.8. - **Same ordering elsewhere:** The ranking carries through to the other sleep measures. Average sleep latency is 46.1 min for TikTok / Reels, 42.5 for Instagram / Reddit and 33.2 for News / Reading. Average morning alarm snoozes are 3.01 for TikTok / Reels and 2.39 for News / Reading. - **Sleep duration barely moves:** Average total sleep is 6.17 h for TikTok / Reels and 6.21 h for Instagram / Reddit, against 6.38 h for News / Reading. The fatigue gap is therefore mostly about how long it takes to fall asleep, not how long people sleep. - **Not driven by other behaviours:** Average post-5pm caffeine (about 32–35 mg), screen brightness (about 55%) and physical activity (about 35–36 min) are nearly identical across apps. - **Age differs, though:** TikTok / Reels users are youngest (average age 31.1) and News / Reading users oldest (38.0). The fatigue gap may partly reflect age, and this is observational data, so it does not show that the app causes the fatigue. **Practical reading:** short-form social video (TikTok / Reels) and Instagram / Reddit users have the highest fatigue and longest sleep latency, so they are the strongest segments to target first. TikTok / Reels is also the largest group, with 2,198 of 8,500 users. {"action":"final","content":"placeholder"} </br>