Comments on: Bay Area R User Group 2009 Kickoff Video http://www.lecturemaker.com/2009/02/r-kickoff-video/ Video marketing for extraordinary people, brands, and products. Tue, 09 Mar 2010 08:09:38 -0700 http://wordpress.org/?v=2.9.1 hourly 1 By: | LectureMaker, LLC http://www.lecturemaker.com/2009/02/r-kickoff-video/comment-page-1/#comment-584 | LectureMaker, LLC Mon, 14 Sep 2009 07:00:14 +0000 http://www.lecturemaker.com/2009/02/r-and-science-of-predictive-analytics-lecture-production/#comment-584 [...] The Science of Predictive Analytics [...] [...] The Science of Predictive Analytics [...]

]]> By: How Google and Facebook are using R :Health Fitness Wealth http://www.lecturemaker.com/2009/02/r-kickoff-video/comment-page-1/#comment-189 How Google and Facebook are using R :Health Fitness Wealth Sat, 18 Jul 2009 22:31:24 +0000 http://www.lecturemaker.com/2009/02/r-and-science-of-predictive-analytics-lecture-production/#comment-189 [...] (March 26th Update: Video now available) Last night, I moderated our Bay Area R Users Group kick-off event with a panel discussion entitled “The R and Science of Predictive Analytics”, co-located with the Predictive Analytics World conference here in SF. [...] [...] (March 26th Update: Video now available) Last night, I moderated our Bay Area R Users Group kick-off event with a panel discussion entitled “The R and Science of Predictive Analytics”, co-located with the Predictive Analytics World conference here in SF. [...]

]]> By: Crossing Arbogast http://www.lecturemaker.com/2009/02/r-kickoff-video/comment-page-1/#comment-186 Crossing Arbogast Thu, 09 Jul 2009 06:30:22 +0000 http://www.lecturemaker.com/2009/02/r-and-science-of-predictive-analytics-lecture-production/#comment-186 Thanks for this Nice post, Really usefull all of us. just bookmarked this post in my digg profile, hope you will update more post soon. I really liked your blog! Regards, Shaza Thanks for this Nice post, Really usefull all of us. just bookmarked this post in my digg profile, hope you will update more post soon.
I really liked your blog!

Regards,
Shaza

]]> By: Video Rig « LectureMaker http://www.lecturemaker.com/2009/02/r-kickoff-video/comment-page-1/#comment-120 Video Rig « LectureMaker Tue, 19 May 2009 09:33:40 +0000 http://www.lecturemaker.com/2009/02/r-and-science-of-predictive-analytics-lecture-production/#comment-120 [...] Link to this video [...] [...] Link to this video [...]

]]> By: Embedded Components and Tools Blog Center » Blog Archive » Bay Area R User Group 2009 Kickoff Video http://www.lecturemaker.com/2009/02/r-kickoff-video/comment-page-1/#comment-66 Embedded Components and Tools Blog Center » Blog Archive » Bay Area R User Group 2009 Kickoff Video Thu, 02 Apr 2009 18:41:08 +0000 http://www.lecturemaker.com/2009/02/r-and-science-of-predictive-analytics-lecture-production/#comment-66 [...] Learn more and watch lecture [...] [...] Learn more and watch lecture [...]

]]> By: Michael Batty http://www.lecturemaker.com/2009/02/r-kickoff-video/comment-page-1/#comment-60 Michael Batty Sun, 29 Mar 2009 22:01:25 +0000 http://www.lecturemaker.com/2009/02/r-and-science-of-predictive-analytics-lecture-production/#comment-60 very interesting here is some other stuff we have done Mike <a href="http://www.nature.com/nature/journal/v444/n7119/abs/nature05302.html" rel="nofollow">Rank clocks</a> <a href="http://www.springerlink.com/content/t42871v22396j04q/" rel="nofollow">Hierarchy in Cities and City Systems</a> very interesting

here is some other stuff we have done

Mike

Rank clocks

Hierarchy in Cities and City Systems

]]> By: admin http://www.lecturemaker.com/2009/02/r-kickoff-video/comment-page-1/#comment-58 admin Sat, 28 Mar 2009 19:45:05 +0000 http://www.lecturemaker.com/2009/02/r-and-science-of-predictive-analytics-lecture-production/#comment-58 I was particularly impressed with how common the <em>power law distribution</em> is used to characterize and predict online user behavior within a well defined ecosystem such as facebook. <strong>power law distribution</strong> [pmath]P(x)=Kx^alpha[/pmath], Where: [pmath]P(x) right [/pmath] size of an event, such as users who stay on facebook more than three weeks [pmath]x right [/pmath] property of an event, such as number of users who join facebook per day [pmath]alpha right [/pmath] some parameter of the distribution, such as the percent of users who enter their gender during new user registration [pmath]K right [/pmath] a scaling constant So the size of a community event, [pmath]P(x)[/pmath], can be predicted with the power law distribution. The problem might reduce down to the selection of the right parameter [pmath]alpha[/pmath] and scaling constant [pmath]K[/pmath] to apply to the properties characterized by [pmath]x[/pmath]. The correlation measured from historical data can then be used to determine if further on-site changes could be made to improve performance from an online community. R then becomes the statistics tool to crunch data, calculate, and plot distributions. Then the results from R can be used to predict and improve future online behavior with new community policies or a change in applications. Neat. Management of large amounts of data was discussed throughout the lecture. One solution is in the open-source cloud initiative called Hadoop. Notes: <ol> <li>LectureMaker supports math publishing, both in its blog posts and in its comments. See my wordpress plugin to learn more, called <a href="http://wordpress.org/extend/plugins/wpmathpub/" rel="nofollow">wpmathpub</a></li> <li>Itamar made reference to <a href="http://hadoop.apache.org/core/" rel="nofollow">Hadoop</a> as one way the facebook data team manages large amounts of data for R projects </li><li>I had to look up "what is a power law distribution" as it applies to social networks. Here is the reference I choose to rely upon <a href="http://www.isoc.org/inet2000/cdproceedings/2a/2a_2.htm#s7" rel="nofollow">Power Law Distributions in Real and Virtual Worlds</a> by Narushige SHIODE and Michael BATTY, University College London </li> </ol> I was particularly impressed with how common the power law distribution is used to characterize and predict online user behavior within a well defined ecosystem such as facebook.

power law distribution

P(x)=Kx^alpha,

Where:

P(x) right size of an event, such as users who stay on facebook more than three weeks
x right property of an event, such as number of users who join facebook per day
alpha right some parameter of the distribution, such as the percent of users who enter their gender during new user registration
K right a scaling constant

So the size of a community event, P(x), can be predicted with the power law distribution. The problem might reduce down to the selection of the right parameter alpha and scaling constant K to apply to the properties characterized by x. The correlation measured from historical data can then be used to determine if further on-site changes could be made to improve performance from an online community.

R then becomes the statistics tool to crunch data, calculate, and plot distributions. Then the results from R can be used to predict and improve future online behavior with new community policies or a change in applications. Neat.

Management of large amounts of data was discussed throughout the lecture. One solution is in the open-source cloud initiative called Hadoop.

Notes:

  1. LectureMaker supports math publishing, both in its blog posts and in its comments. See my wordpress plugin to learn more, called wpmathpub
  2. Itamar made reference to Hadoop as one way the facebook data team manages large amounts of data for R projects
  3. I had to look up “what is a power law distribution” as it applies to social networks. Here is the reference I choose to rely upon Power Law Distributions in Real and Virtual Worlds by Narushige SHIODE and Michael BATTY, University College London
]]> By: How Google and Facebook are using R : Data Evolution http://www.lecturemaker.com/2009/02/r-kickoff-video/comment-page-1/#comment-57 How Google and Facebook are using R : Data Evolution Thu, 26 Mar 2009 10:17:00 +0000 http://www.lecturemaker.com/2009/02/r-and-science-of-predictive-analytics-lecture-production/#comment-57 [...] Update: Video now available! Last night, I moderated our Bay Area R Users Group kick-off event with a panel discussion entitled “The R and Science of Predictive Analytics”, co-located with the Predictive Analytics World conference here in SF. [...] [...] Update: Video now available! Last night, I moderated our Bay Area R Users Group kick-off event with a panel discussion entitled “The R and Science of Predictive Analytics”, co-located with the Predictive Analytics World conference here in SF. [...]

]]> By: Michael E. Driscoll http://www.lecturemaker.com/2009/02/r-kickoff-video/comment-page-1/#comment-56 Michael E. Driscoll Thu, 26 Mar 2009 10:12:20 +0000 http://www.lecturemaker.com/2009/02/r-and-science-of-predictive-analytics-lecture-production/#comment-56 Hi Ron - Many thanks for putting this together - the video looks great! I particularly like the way you've allowed browsing of the content with the red status bars below. Look forward to more great stuff in the future. Mike Hi Ron – Many thanks for putting this together – the video looks great! I particularly like the way you’ve allowed browsing of the content with the red status bars below. Look forward to more great stuff in the future.

Mike

]]>