Ask anything, I'll see what I can answer. On Feb 11, 2015, at 12:47 PM, Richard Newman <[email protected]> wrote: > * What are the 95% or 98% sizes for upload sizes and locations? Interested in > the extremes.
The upper ~5% is 100-250 kb. 250 is the bounded upper limit on stored data. http://mzl.la/1EYAlOm The upper tail on the number of GPS locations is probably best seen visually: http://mzl.la/1EYAVvo Anyone running a tracking/navigation app (like Strava, Google Maps navigation), will get hundreds or thousands of GPS locations. > * Do we have enough info to partition our users into ones that capture and > ones that don't? Presumably we don't report zero locations captured, zero > bytes uploaded. I'm interested to see what the contribution ratio is. Interesting question, we don't have this metric, I think we should add this. The telemetry has mainly been focused on proving we aren't burdening users. > * Does the MLS distinguish between passive and active stumbles? If so, are we > able to produce two maps with traces to visually compare? Not as a map. There are metrics on input data volumes (Fennec vs. Github stumbler vs. Play Store stumbler vs. other sources). There is data available on the contribution value (such as new cell discoveries, wifi discoveries) per contribution source. One thing a map would show, which we are still gathering data on, is the value per-region of these different stumbling methods. The Mozilla Stumbler app has a presence in only a handful of countries -we haven't been able to get the message out beyond North America + Western Europe- so it would be interesting to see in what new regions Fennec stumbling is gathering data. > On Wed, Feb 11, 2015 at 9:01 AM, Garvan Keeley <[email protected]> wrote: > To get an idea of the impact this feature has had, here are some stats. > > * Due to Fennec user contributions, MLS has collected as much Wi-Fi and cell > data in the last 6 weeks as it has in 6 months of 2014 > > The per-user burden is minimal, as indicated by the following stats: > > Passively received GPS locations per day: > * 1-10 per day ~50% > * 10-100 per day ~40% > * 100-2500 per day ~ 10% > > Upload costs > * 90% of uploads are less than 15kb > * 2/3 of uploads are 24 hours apart or greater > > One concern we had was that most users' devices were not going to trigger the > GPS enough for us to collect data by passively listening for the GPS. So far, > this does not appear to be the case. > _______________________________________________ > mobile-firefox-dev mailing list > [email protected] > https://mail.mozilla.org/listinfo/mobile-firefox-dev > _______________________________________________ mobile-firefox-dev mailing list [email protected] https://mail.mozilla.org/listinfo/mobile-firefox-dev
