Hyndsight

 

A blog by Rob J Hyndman 

Twitter Gplus RSS

Stack exchange for statistical analysis needs you!

Published on 18 June 2010

The pro­posal to cre­ate a Stack­Ex­change site for sta­tis­ti­cal analy­sis is steadily mov­ing for­ward. We have now com­pleted the scop­ing stage which involved find­ing enough peo­ple will­ing to express an inter­est in the idea, and vot­ing on some exam­ple ques­tions to define what is allowed and what is not allowed on the site. The on-​​topic ques­tions that have been selected are these:

  1. What is a ‘stan­dard deviation’?
  2. Which of the fol­low­ing three graph­ics best dis­plays this data set? Why?
  3. What’s the best way to iden­tify an out­lier in mul­ti­vari­ate data?
  4. Can you give an exam­ple of where I might pre­fer to use a z-​​test vs a t-​​test?
  5. What are the dif­fer­ences between Bayesian and Fre­quen­tist reasoning?

Exam­ples of ques­tions con­sid­ered off-​​topic are:

  1. How do I win in Poker?
  2. I have two chil­dren. One is a boy born on a Tues­day. What is the prob­a­bil­ity I have two boys?
  3. Joe is 8 years old, Mike is 10 years old, and Alice is 13. What is their MEDIAN age?
  4. Where can I access NASA’s data archives?
  5. How much should I expect to pay for a SAS licence?

The next phase is to get peo­ple to com­mit to con­tribut­ing to the site. Many read­ers of this blog have already reg­is­tered as “fol­low­ers” — now you have to make a com­mit­ment to be a con­trib­u­tor as well. The site won’t launch until there are enough peo­ple com­mit­ted to being part of it.

Just go to the site and indi­cate that you are will­ing to be an active par­tic­i­pant once it launches.

If you’re won­der­ing what this is all about, and why this is a much bet­ter approach than the var­i­ous usenet and email help groups, there’s a nice sum­mary on Tal Galili’s blog.

Related Posts:

 
Tags: , , ,
13 Comments  comments 
  • http://www.r-statistics.com Tal Galili

    Hi Rob,
    Thanks for link­ing, and espe­cially for all the amaz­ing work you have put into mak­ing this site go online.
    I am look­ing for­ward to look­ing at this stage a year from now and think­ing “how could it even take so long to start”.

    Great job!
    Tal

  • http://www.r-statistics.com Tal Galili

    Hi Rob,
    Also,
    Please tag this post with “R”, so it will get repub­lished on R-​​bloggers!

    Cheers,
    Tal

  • Faiz Rasool

    Very nice idea. With increas­ing use of sta­tis­tics in social sci­ences, and in other fields, the num­ber of peo­ple who want to learn sta­tis­tics has risen. At the same time, many stu­dents find it dif­fi­cult to get sound answers to their basic yet highly sig­nif­i­cant ques­tions. I hope this idea trans­lates into a suc­cess­ful website.

  • http://www.statisticsblog.com Matt Asher

    I would *very strongly* sug­gest that you build the site using one of the free ver­sions of Q&A soft­ware that com­pete with stack­ex­change. Not only is stack­ex­change a walled gar­den con­trolled by another com­pany, but “Joel and Jeff”, the cre­ators, have already shown that impos­ing their vision is more impor­tant than respect­ing their cus­tomers. Before going for­ward with their plat­form make sure to read the his­tory of how SE was imple­mented, and think hard about why they decided cus­tomers wouldn’t even be allowed to down­load full data on their own users, even those cus­tomers who they were uni­lat­er­ally shut­ting down. I never cre­ated a site on their plat­form myself, but I con­sid­ered it very care­fully and now I’m glad I didn’t.

    At any rate this is clearly an idea whose time has come, and I would link to this right away. Just make sure to be very care­ful about whose plat­form you build it on.

    Cheers,

    Matt

    • http://robjhyndman.com Rob J Hyndman

      Hi Matt. I am aware of the his­tory of Stack­Ex­change — I am a mod­er­a­tor on a SE 1.0 site. I think they have made a cou­ple of mis­takes, but I wouldn’t con­demn them for that. On the other hand, the free Q&A soft­ware options are largely rip-​​offs of the SE idea, and steal­ing IP is not nice.

      By the way, hand­ing out user’s data to third par­ties is ille­gal in this coun­try, so that could be why they have been unwill­ing to allow full data download.

      I thought about the other options before propos­ing the SE site, and decided it was bet­ter to go with a group who were pas­sion­ate about mak­ing excel­lent soft­ware, even if we dis­agree with some of their choices, then to use look-​​alike soft­ware that might not work so well.

      • http://www.osqa.net/ Rick Ross

        Hi Rob,

        I can look past you dis­re­gard­ing the fact that the Stack­Ya­daYada sites, them­selves, are hugely deriv­a­tive works. To sug­gest that the devel­op­ers of FOSS solu­tions like OSQA, Sha­pado, and oth­ers are not pas­sion­ate about devel­op­ing excel­lent soft­ware is sim­ply incorrect.

        We will gladly place our com­mit­ment to the com­mu­nity and our track record of respon­sive inno­va­tion up against any­one in the field. OSQA is not try­ing to “catch up” to Stack­Ya­daYada — we are accel­er­at­ing past them quickly. OSQA not only works well, but increas­ingly it does things no other Q&A com­mu­nity sys­tem can do. You should try it before you judge it.

        Cheers,
        Rick

        • http://robjhyndman.com Rob J Hyndman

          Hi Rick. I’m sorry if I mis-​​represented the sit­u­a­tion regard­ing SE and the free alter­na­tives. Best wishes, Rob

      • http://www.statisticsblog.com Matt Asher

        Hi Rob,

        I under­stand this may now be a fait accom­pli, but if you haven’t seen it read Joel’s recent blog post on domain names:

        http://​blog​.stack​over​flow​.com/​2​0​1​0​/​0​7​/​d​o​m​a​i​n​-​n​ames/

        From the quote at the end about them “help­ing out” with the domain name it should be crys­tal clear what’s going on. Stack­Ex­change is a grab for not only all con­tent gen­er­ated by the effort of you and your users, but now he’s wran­gling for con­trol of people’s domain names as well.

        Would you build an office on leased land, using leased bricks? Would you let some other com­pany con­trol your cus­tomer data­base and limit how you access it? This is not about ensur­ing pri­vacy, it’s about con­trol and lock-​​in. Look at how they posted info about traf­fic lev­els to all mem­ber sites under the orig­i­nal pro­gram. No doubt that’s info those sites (some pay­ing cus­tomers!) would have pre­ferred to keep private.

        I wish you the best but watch out.

        Cheers,

        Matt

        • http://robjhyndman.com Rob J Hyndman

          I fail to see why the SE folk gen­er­ate such strong reac­tions. I’ve read the blog arti­cle and it seemed fairly sen­si­ble and harm­less to me. The sta­tis­ti­cal analy­sis site is not going to be *mine*, it will be a com­mu­nity site. The par­tic­i­pants are not cus­tomers, but other mem­bers of the com­mu­nity. And I’m very happy that SE are will­ing to pro­vide the plat­form and infra­struc­ture to host it.

          I must be miss­ing some­thing, but there seem to be a lot of peo­ple with strong opin­ions on this.

  • Manuel Ramón

    Hello Rob
    I have joined the Sta­tis­ti­cal Analy­sis site but I can’t find how to pro­pose new ques­tions or how to vote the exist­ing one. Can you help me?

    • http://robjhyndman.com Rob J Hyndman

      Hi Manuel. The site is now in the com­mit­ment phase which means it is no longer accept­ing new ques­tions or votes. Once we have enough com­mit­ted users, it will move into the beta phase when I imag­ine gen­uine ques­tions and answers can be added. The best thing to do dur­ing the com­mit­ment phase is to encour­age every­one you know who might be inter­ested to join up.

  • Pingback: “Statistical Analysis Q&A website” did StackOverFlow just lose it to MetaOptimize? and is it good or bad? | R-statistics blog

  • Pingback: StackOverFlow and MetaOptimize are battling to be the #1 “Statistical Analysis Q&A website” – to whom would you signup? | sumber referensi statistika