The ol' favoriting reacharound. March 1, 2010 4:01 PM Subscribe
Some specific MeFi info-mashing about favorites I haven't seen.
It'll probably be easiest to tell you what I'm interested in since I have none of the knowledge necessary to make such a query.
Has anyone done anything with infodumps that would track stuff like how many times one user has favorited another's comments, or favorited/commented on another's posts?
I'd been wondering this for a while, but without forming the thought so clearly, I thought it might be interpreted as kind of stalkerish. I don't think so, though. I feel like I've noticed favoriting or being favorited by particular users multiple times, but I have no way of knowing if this is confirmation bias for memorable usernames, or some users favoriting everything in sight or what.
It'll probably be easiest to tell you what I'm interested in since I have none of the knowledge necessary to make such a query.
Has anyone done anything with infodumps that would track stuff like how many times one user has favorited another's comments, or favorited/commented on another's posts?
I'd been wondering this for a while, but without forming the thought so clearly, I thought it might be interpreted as kind of stalkerish. I don't think so, though. I feel like I've noticed favoriting or being favorited by particular users multiple times, but I have no way of knowing if this is confirmation bias for memorable usernames, or some users favoriting everything in sight or what.
A lot of this went on in the Infodump 2.0 thread, starting with this comment by FishBike (who did most of the number-crunching in this regard).
posted by Rhaomi at 4:11 PM on March 1, 2010
posted by Rhaomi at 4:11 PM on March 1, 2010
This sort of thing has come up a few times before in one way or another (see previous threads tagged with 'infodump'); FishBike has a tool that calculates a number of different favorites-interaction and mutual-favoriting tables for an individual user, so you could try mefimailing him if he doesn't show up here (though I bet he'll show up here).
There's no on-site facility for calculating this info, though if you do decide to try and explore the data yourself you can use the files in the Infodump, though I gather from your tags that you're already aware it exists.
posted by cortex (staff) at 4:12 PM on March 1, 2010
There's no on-site facility for calculating this info, though if you do decide to try and explore the data yourself you can use the files in the Infodump, though I gather from your tags that you're already aware it exists.
posted by cortex (staff) at 4:12 PM on March 1, 2010
So, in summary: infodump, Fishbike, no. Miller time!
posted by cortex (staff) at 4:13 PM on March 1, 2010
posted by cortex (staff) at 4:13 PM on March 1, 2010
Yes, FishBike has a script that will do this.
Yep.
I believe he'll run the numbers for folks on request,
Yep. MeFiMail welcome from anyone interested. I seem to get 10-20 requests every time this comes up, which is cool.
but only the requester's data not other users.
Yep. It's a little too much scrutiny of one person's activity to run it without their consent, I think.
(though I bet he'll show up here)
Nope. Oh wait... by the time anybody reads this, yep.
cmoj, I'll send you the current version of the analysis for your account. I've been making a few minor changes to it recently, so this'll get me to finish those off. If it's not quite what you had in mind, let's discuss it here because any excuse for a datawankery thread will do.
posted by FishBike at 4:20 PM on March 1, 2010
Yep.
I believe he'll run the numbers for folks on request,
Yep. MeFiMail welcome from anyone interested. I seem to get 10-20 requests every time this comes up, which is cool.
but only the requester's data not other users.
Yep. It's a little too much scrutiny of one person's activity to run it without their consent, I think.
(though I bet he'll show up here)
Nope. Oh wait... by the time anybody reads this, yep.
cmoj, I'll send you the current version of the analysis for your account. I've been making a few minor changes to it recently, so this'll get me to finish those off. If it's not quite what you had in mind, let's discuss it here because any excuse for a datawankery thread will do.
posted by FishBike at 4:20 PM on March 1, 2010
Hey FB you can run my numbers and post them here if you want, just to give people an idea of what that sort of thing looks like.
posted by jessamyn (staff) at 4:38 PM on March 1, 2010
posted by jessamyn (staff) at 4:38 PM on March 1, 2010
Thanks, jessamyn!
I've been altering the script slightly, taking out a couple of sections that really weren't interesting, and adding in a few more over time. I still haven't figured out exactly how to do something with the tag data and user activity, without mentioning specific tags in the report, so that's still in the works.
Stats for:jessamyn (v1.1.0 beta 2)
posted by FishBike at 4:53 PM on March 1, 2010 [6 favorites]
I've been altering the script slightly, taking out a couple of sections that really weren't interesting, and adding in a few more over time. I still haven't figured out exactly how to do something with the tag data and user activity, without mentioning specific tags in the report, so that's still in the works.
Stats for:jessamyn (v1.1.0 beta 2)
Who does jessamyn favorite the most?
(simple count of favorites)
not_on_display [74]
cortex [36]
Greg Nog [19]
Astro Zombie [15]
dersins [13]
Brandon Blatcher [12]
Ambrosia Voyeur [10]
melissa may [9]
klangklangston [9]
BitterOldPunk [8]
Who does jessamyn favorite the most?
(percent of their comments+posts since you joined)
(limited to users you've favorited 5+ times)
1.72% (74 of 4295) of not_on_display's comments+posts
1.13% (9 of 794) of melissa may's comments+posts
0.67% (19 of 2831) of Greg Nog's comments+posts
0.62% (5 of 803) of Skot's comments+posts
0.51% (5 of 979) of FauxScot's comments+posts
0.42% (8 of 1922) of Baby_Balrog's comments+posts
0.32% (5 of 1578) of taz's comments+posts
0.30% (8 of 2662) of BitterOldPunk's comments+posts
0.25% (7 of 2825) of ColdChef's comments+posts
0.22% (8 of 3720) of Meatbomb's comments+posts
Who gets the highest percentage of their favorites from jessamyn?
(limited to users you've favorited 5+ times)
not_on_display [1.561%: 74 of 4742 favorites]
FauxScot [0.967%: 5 of 517 favorites]
melissa may [0.428%: 9 of 2101 favorites]
iconomy [0.372%: 6 of 1611 favorites]
Skot [0.312%: 5 of 1602 favorites]
Baby_Balrog [0.233%: 8 of 3429 favorites]
Meatbomb [0.223%: 8 of 3595 favorites]
Deathalicious [0.214%: 5 of 2336 favorites]
vacapinta [0.212%: 5 of 2358 favorites]
taz [0.203%: 5 of 2462 favorites]
Who favorites jessamyn the most?
(simple count of favorites)
koeselitz [291]
scrump [281]
deborah [232]
not_on_display [187]
bru [169]
tehloki [145]
nooneyouknow [145]
scody [142]
DevilsAdvocate [131]
shmegegge [130]
Who favorites jessamyn the most?
(percent of your comments+posts since they joined)
(as long as you've made 50+ comments+posts since they joined)
bakerina: 3.55% (55 of 1549) of jessamyn's comments+posts
koeselitz: 2.07% (291 of 14074) of jessamyn's comments+posts
scrump: 2.00% (281 of 14074) of jessamyn's comments+posts
not_on_display: 1.92% (187 of 9745) of jessamyn's comments+posts
deborah: 1.65% (232 of 14074) of jessamyn's comments+posts
burnmp3s: 1.32% (124 of 9381) of jessamyn's comments+posts
bru: 1.20% (169 of 14074) of jessamyn's comments+posts
PhoBWanKenobi: 1.20% (79 of 6607) of jessamyn's comments+posts
tehloki: 1.15% (145 of 12643) of jessamyn's comments+posts
Pope Guilty: 1.09% (126 of 11598) of jessamyn's comments+posts
Who favorites jessamyn the most?
(percentage of their total favorites)
(limited to users who've favorited you 5+ times)
judith [25.000%: 5 of 20 favorites]
mrmorgan [23.810%: 5 of 21 favorites]
bru [23.151%: 169 of 730 favorites]
terrapin [8.911%: 9 of 101 favorites]
SlyBevel [8.171%: 21 of 257 favorites]
donnagirl [6.658%: 53 of 796 favorites]
PugAchev [6.612%: 8 of 121 favorites]
beefetish [6.452%: 8 of 124 favorites]
ersatzkat [6.215%: 44 of 708 favorites]
ottereroticist [6.048%: 15 of 248 favorites]
Who has favorited the same items as jessamyn the most?
(by simple count of mutual favorites)
scrump [87]
koeselitz [72]
deborah [69]
misha [63]
schyler523 [60]
shmegegge [57]
tehloki [54]
burnmp3s [49]
Caduceus [47]
flibbertigibbet [47]
Who has favorited the same items as jessamyn the most?
(by percentage of other user's favorites you also favorited)
(limited to users with 5+ mutual favorites)
jacobian [7.06%] (6 of 85 favorites)
explosion [6.94%] (5 of 72 favorites)
lola [6.41%] (5 of 78 favorites)
cbp [5.95%] (5 of 84 favorites)
Mo Nickels [5.88%] (5 of 85 favorites)
davros42 [5.75%] (5 of 87 favorites)
goo [5.62%] (5 of 89 favorites)
pseudonick [5.61%] (6 of 107 favorites)
JeremiahBritt [4.95%] (5 of 101 favorites)
chowflap [4.52%] (10 of 221 favorites)
Of the threads where jessamyn has been active, who else has been active in the highest percentage?
(limited to threads active after the comparison user has joined MetaFilter)
cortex: 23.0% [2655 of 11563]
languagehat: 16.8% [1899 of 11287]
mathowie: 14.9% [1734 of 11655]
Brandon Blatcher: 14.8% [1521 of 10299]
klangklangston: 13.9% [1360 of 9789]
Alvy Ampersand: 12.7% [1132 of 8901]
Burhanistan: 12.6% [883 of 7017]
kathrineg: 12.4% [226 of 1827]
stavrosthewonderchicken: 11.8% [1368 of 11607]
delmoi: 11.2% [1287 of 11502]
Of the threads where other users have been active, in whose has jessamyn also been the most active by percentage?
(limited to threads active after jessamyn has joined MetaFilter)
the Cabal: 69.8% [37 of 53]
and hosted from Uranus: 43.9% [272 of 620]
double block and bleed: 43.8% [99 of 226]
Kwine: 39.5% [386 of 978]
Ceiling Cat: 37.7% [20 of 53]
palomar: 37.2% [29 of 78]
Duncan: 35.8% [54 of 151]
CKmtl: 35.5% [366 of 1030]
waraw: 35.1% [185 of 527]
subbes: 35.0% [123 of 351]
posted by FishBike at 4:53 PM on March 1, 2010 [6 favorites]
Someone's love is not on display.
posted by gman at 5:24 PM on March 1, 2010 [4 favorites]
posted by gman at 5:24 PM on March 1, 2010 [4 favorites]
Who does jessamyn favorite the most?
(simple count of favorites)
not_on_display [74]
But does he favourite her the most in return?
posted by crossoverman at 5:25 PM on March 1, 2010
(simple count of favorites)
not_on_display [74]
But does he favourite her the most in return?
posted by crossoverman at 5:25 PM on March 1, 2010
you're ruining everything.
posted by gman at 5:30 PM on March 1, 2010 [1 favorite]
posted by gman at 5:30 PM on March 1, 2010 [1 favorite]
crossoverman:
Who favorites jessamyn the most?
(percent of your comments+posts since they joined)
(as long as you've made 50+ comments+posts since they joined)
bakerina: 3.55% (55 of 1549) of jessamyn's comments+posts
koeselitz: 2.07% (291 of 14074) of jessamyn's comments+posts
scrump: 2.00% (281 of 14074) of jessamyn's comments+posts
not_on_display: 1.92% (187 of 9745) of jessamyn's comments+posts
posted by jacalata at 6:05 PM on March 1, 2010
Who favorites jessamyn the most?
(percent of your comments+posts since they joined)
(as long as you've made 50+ comments+posts since they joined)
bakerina: 3.55% (55 of 1549) of jessamyn's comments+posts
koeselitz: 2.07% (291 of 14074) of jessamyn's comments+posts
scrump: 2.00% (281 of 14074) of jessamyn's comments+posts
not_on_display: 1.92% (187 of 9745) of jessamyn's comments+posts
posted by jacalata at 6:05 PM on March 1, 2010
GRAR FAVORITES
posted by brain_drain at 6:15 PM on March 1, 2010
posted by brain_drain at 6:15 PM on March 1, 2010
But does he favourite her the most in return?
There are two ways to interpret that question. One is "does anyone favorite you more than user X?" and in this case we can see that three users do. The other interpretation is "does user X favorite anyone more than they favorite you", and that information isn't in the current version of the stats report.
Maybe it could be, though, something like a "who favorites you the most, by position on their most-favorited list", and you'd get to know who the people are who favorite you more than they favorite anyone else.
I don't know though, is that too much focus on the user X's rather than the person the stats are for?
posted by FishBike at 6:17 PM on March 1, 2010
There are two ways to interpret that question. One is "does anyone favorite you more than user X?" and in this case we can see that three users do. The other interpretation is "does user X favorite anyone more than they favorite you", and that information isn't in the current version of the stats report.
Maybe it could be, though, something like a "who favorites you the most, by position on their most-favorited list", and you'd get to know who the people are who favorite you more than they favorite anyone else.
I don't know though, is that too much focus on the user X's rather than the person the stats are for?
posted by FishBike at 6:17 PM on March 1, 2010
I heard you can sacrifice 10 most-favoriters for a double whopper.
posted by qvantamon at 6:21 PM on March 1, 2010
posted by qvantamon at 6:21 PM on March 1, 2010
Wow. That's...impressive.
posted by (Arsenio) Hall and (Warren) Oates at 6:39 PM on March 1, 2010
posted by (Arsenio) Hall and (Warren) Oates at 6:39 PM on March 1, 2010
The stats script isn't published anywhere, yet. There are a bunch of reasons why not, but I suspect the biggest is that I selfishly feel like it's my pony and I'm not done playing with him yet. If I ever get bored with running stats for people, I'm going to have to figure out what to do with the code.
Anyway, the script is just a bunch of fairly poorly written1 SQL queries, and I do have a page here with some code people can use to import the Infodump into an MS SQL or MySQL database easily if folks want to mess around with it themselves. There's even some code there to reformat the data a little, to make it easier to query.
1: the poorly written nature being another reason why they're a secret... a fair bit of work would be needed before I'd let anybody see the code, sort of like tidying up the place before having guests, even though I'm sure their place is normally as bad as mine is.
posted by FishBike at 7:02 PM on March 1, 2010 [2 favorites]
Anyway, the script is just a bunch of fairly poorly written1 SQL queries, and I do have a page here with some code people can use to import the Infodump into an MS SQL or MySQL database easily if folks want to mess around with it themselves. There's even some code there to reformat the data a little, to make it easier to query.
1: the poorly written nature being another reason why they're a secret... a fair bit of work would be needed before I'd let anybody see the code, sort of like tidying up the place before having guests, even though I'm sure their place is normally as bad as mine is.
posted by FishBike at 7:02 PM on March 1, 2010 [2 favorites]
Oh dear god. Now I'm gonna feel really weird every time I want to favourite something of jessamyn's.
posted by deborah at 7:12 PM on March 1, 2010
posted by deborah at 7:12 PM on March 1, 2010
It might be interesting to graph some of the network effects.
posted by five fresh fish at 7:15 PM on March 1, 2010
posted by five fresh fish at 7:15 PM on March 1, 2010
I would like to be able to click on someone's profile and see "This user has favorited you X times."
posted by desjardins at 7:30 PM on March 1, 2010 [1 favorite]
posted by desjardins at 7:30 PM on March 1, 2010 [1 favorite]
jessamyn (v1.1.0 beta 2)
Beta 2?? Come on, dude. You really need to be loading the nightly mod builds. Sure, they're a little flaky, but you're lucky the any of the .0 betas even load! (Did I mention the oversized ban hammer that was added in r55304? So. Great.)
posted by niles at 8:26 PM on March 1, 2010
Beta 2?? Come on, dude. You really need to be loading the nightly mod builds. Sure, they're a little flaky, but you're lucky the any of the .0 betas even load! (Did I mention the oversized ban hammer that was added in r55304? So. Great.)
posted by niles at 8:26 PM on March 1, 2010
Fishbike, if you're not already swamped would you please run my numbers? I'm perfectly fine with them being posted here if so, but if you prefer you can memail me.
posted by Solon and Thanks at 8:31 PM on March 1, 2010
posted by Solon and Thanks at 8:31 PM on March 1, 2010
You really need to be loading the nightly mod builds. Sure, they're a little flaky
oh man remember when we rolled out the cortex 0.41.4 build and I'd accidentally inverted the conditional logic on the doBanhammer function and everybody who didn't self-link that day got locked out until pb could roll back the changes but pb couldn't do that because he hadn't self-linked either man that was great
posted by cortex (staff) at 8:40 PM on March 1, 2010 [2 favorites]
oh man remember when we rolled out the cortex 0.41.4 build and I'd accidentally inverted the conditional logic on the doBanhammer function and everybody who didn't self-link that day got locked out until pb could roll back the changes but pb couldn't do that because he hadn't self-linked either man that was great
posted by cortex (staff) at 8:40 PM on March 1, 2010 [2 favorites]
Of the threads where jessamyn has been active, who else has been active in the highest percentage?
Also known as the You Spend Too Much Time on MeTa list.
posted by Alvy Ampersand at 8:56 PM on March 1, 2010
Also known as the You Spend Too Much Time on MeTa list.
posted by Alvy Ampersand at 8:56 PM on March 1, 2010
The Cabal spends too much time on MeTa? That's almost tautological.
posted by UbuRoivas at 9:28 PM on March 1, 2010
posted by UbuRoivas at 9:28 PM on March 1, 2010
I do have a page here with some code people can use to import the Infodump into an MS SQL or MySQL database easily if folks want to mess around with it themselves.
This is awesome.
posted by A Terrible Llama at 4:05 AM on March 2, 2010
This is awesome.
posted by A Terrible Llama at 4:05 AM on March 2, 2010
Oh dear god. Now I'm gonna feel really weird every time I want to favourite something of jessamyn's.
Yeah, I learned I had an unnatural affinity for what cortex writes.
posted by cjorgensen at 6:28 AM on March 2, 2010
Yeah, I learned I had an unnatural affinity for what cortex writes.
posted by cjorgensen at 6:28 AM on March 2, 2010
Of the threads where other users have been active, in whose has jessamyn also been the most active by percentage?
(limited to threads active after jessamyn has joined MetaFilter)
This is the last category in FishBike's analysis. I'm having a hard time parsing it. Can someone break it down for me in more digestible language?
posted by slogger at 6:38 AM on March 2, 2010
(limited to threads active after jessamyn has joined MetaFilter)
This is the last category in FishBike's analysis. I'm having a hard time parsing it. Can someone break it down for me in more digestible language?
posted by slogger at 6:38 AM on March 2, 2010
Ironically, I haven't favorited enough to get any meaningful stats for myself yet. It seems to take about 1000 favorites for the top 10 lists to start looking about right, though even half of that will work kind of OK if I lower the thresholds for inclusion a bit.
At the rate I'm going, it'll be 2014 before I can get a useful report for me.
posted by FishBike at 6:39 AM on March 2, 2010
At the rate I'm going, it'll be 2014 before I can get a useful report for me.
posted by FishBike at 6:39 AM on March 2, 2010
I guess nevermind about my running my stats. I've only favourited a handful.
posted by ODiV at 6:52 AM on March 2, 2010
posted by ODiV at 6:52 AM on March 2, 2010
Of the threads where other users have been active, in whose has jessamyn also been the most active by percentage?
(limited to threads active after jessamyn has joined MetaFilter)
This is the last category in FishBike's analysis. I'm having a hard time parsing it. Can someone break it down for me in more digestible language?
Yeah, that language is a little bit opaque, isn't it? Let me see if I can explain it a little more clearly.
For every user, I'm looking at how many threads they have been active in, where "active" means they're either the poster who started it, or they commented in the discussion. Then I look at, of those threads, in what percentage was jessamyn also active?
So let's say you've been active in 100 threads. Of those, jessamyn has shown up in 10 of them. On this section of the report for jessamyn, we'd see your name with "10%" beside it. This is just a top 10 list of users with the highest percentages calculated that way.
Since we can't expect people to be active in threads that happened before they started here, we're only counting the threads that you've been active in since jessamyn's sign-up date.
Does is make any more sense now?
posted by FishBike at 6:54 AM on March 2, 2010
(limited to threads active after jessamyn has joined MetaFilter)
This is the last category in FishBike's analysis. I'm having a hard time parsing it. Can someone break it down for me in more digestible language?
Yeah, that language is a little bit opaque, isn't it? Let me see if I can explain it a little more clearly.
For every user, I'm looking at how many threads they have been active in, where "active" means they're either the poster who started it, or they commented in the discussion. Then I look at, of those threads, in what percentage was jessamyn also active?
So let's say you've been active in 100 threads. Of those, jessamyn has shown up in 10 of them. On this section of the report for jessamyn, we'd see your name with "10%" beside it. This is just a top 10 list of users with the highest percentages calculated that way.
Since we can't expect people to be active in threads that happened before they started here, we're only counting the threads that you've been active in since jessamyn's sign-up date.
Does is make any more sense now?
posted by FishBike at 6:54 AM on March 2, 2010
You are not logged in, either login or create an account to post comments
posted by pb (staff) at 4:10 PM on March 1, 2010