[00:09:20] (PS1) Milimetric: moving to locally hosted js [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/77261 [00:09:29] (CR) Milimetric: [C: 2 V: 2] moving to locally hosted js [analytics/wikimetrics] - https://gerrit.wikimedia.org/r/77261 (owner: Milimetric) [07:50:00] heya [07:50:25] qchris, amit has requested to interpolate data for the missing days, can you add that and update the dashboards? [10:27:25] (PS1) Jdlrobson: Show graph of first time edits [analytics/limn-mobile-data] - https://gerrit.wikimedia.org/r/77278 [11:38:42] yo qchris [11:38:49] Morning drdee [11:38:57] heya [11:39:49] amit has requested to interpolate data for the missing days, can you add that and update the dashboards? [11:40:01] Yes, I received that email :-( [11:40:06] I am already working on that. [11:40:12] awesome! [11:40:12] ty [11:50:37] drdee: Rereading your line above ... just to align our expectations, I'll not really interpolate /data/ as that would be faking data. But I'll adapt the genterated graphs so each data point is connected to the next, even if they are >1 day apart. [11:50:45] drdee: so something like [11:50:55] drdee: http://gp.wmflabs.org/graphs/free_mobile_traffic_by_version [11:51:28] drdee: (The fix for that link is just temporary. I'll go away in ~9 minutes. So you cannot send that to Amit yet.) [11:51:38] mmmmmm [11:51:49] ok, not sure if i really like this solution but it's good for now [11:52:00] drdee: In that graph, look at the 15th May-23rd May. [11:52:36] drdee: "not sure if i really like this solution" ... any other solution would be faking data. I would not want that. [11:52:44] drdee: Would you like to fake data? [11:54:33] interpolating data is not faking data [11:55:05] If we're selling an interpolated data point for the 20th, for which we have no data point, that is indeed faking data. [11:55:38] We'd be selling each datapoint the same. Regardless if it was a real one or a faked one. [11:56:18] it is replacing missing observations with algorithmically generated data; faking data has a strong connotation that we just put in random numbers. Maybe we should add a feature to limn that would give a brush style to interpolated data so it is visually recognizable [11:57:29] With the current limn, it is already visible, which day ha a real data point. [11:57:43] (dots when hovering a date) [11:58:11] it's good for now :) [17:18:12] ja, average, btw, def ask questions if you ever feel you are out of the loop on this stuff at all [17:18:21] I'm happy to explain and keep everyone updated [19:22:39] ping ottomatata [19:22:39] ping ottomata [19:22:39] ping ottomata1 [19:23:58] piinggeee [19:24:08] drdee