Gaps between plays

You are a product analyst at the music streaming service TuneWave. The recommendations team wants to know how densely listeners consume tracks: short pauses mean continuous playlist listening, long ones mean the person re-opened the app. You have the track-start log play_events: who pressed Play and when. For every event, compute how many minutes passed since the previous event of the same user. Skip the very first event of each user — there is nothing to compare it with. Return columns: - user_id — listener identifier; - played_at — track start time (timestamp, as stored); - minutes_since_prev — whole number of minutes since the user's previous event (all gaps in the data are exact minutes). Sort by user_id ascending, then played_at ascending. Users with a single event do not appear in the output. Timestamps never repeat within one user.