Assess predictive power of the VIDEO_DROPPED_FRAMES_PROPORTION probe
Categories
(Data Science :: Investigation, task)
Tracking
(Not tracked)
People
(Reporter: esmyth, Assigned: ccd)
Details
(Whiteboard: [dsperf])
Brief Description of the request (required):
We believe that a VIDEO_DROPPED_FRAMES_PROPORTION greater than 5% negatively affects user engagement and retention and that the effect is more significant when the rate exceed 10%. Please analyze telemetry from Firefox Desktop nightly and beta between versions 47 and 54 inclusive to assess whether the VIDEO_DROPPED_FRAMES_PROPORTION probe has predictive value for user engagement and retention. Specifically, please consider
- Independent variable: VIDEO_DROPPED_FRAMES_PROPORTION.
- Dependent variables: Retention, session length, active hours, URI count
- Group by: Days, weeks
- Possible segments: OS, memory, cores, VIDEO_CAN_CREATE_H264_DECODER
Business purpose for this request (required):
To produce a metric to guide efforts related to media playback performance and quality of experience.
Requested timelines for the request or how this fits into roadmaps or critical decisions (required):
Rough analysis requested by EOD March 20th. Full analysis by EOD April 5th.
Links to any assets (e.g Start of a PHD, BRD; any document that helps describe the project):
Name of Data Scientist (If Applicable):
Corey Dow-Hygelund
|
||
Updated•7 years ago
|
|
|
Assignee | |
Comment 1•7 years ago
|
||
|
|
Assignee | |
Comment 2•6 years ago
|
||
Media team has made request to make VIDEO_DROPPED_FRAMES_PROPORTION and VIDEO_PLAY_TIME permanent and collected in release. Looks to be riding the FF69 train. This task will be revisited in a couple months, once adequate data has been collected.
|
|
Assignee | |
Comment 4•6 years ago
|
||
Yes, it makes sense to close and adding a new one then.
Description
•