Loading ...
Sorry, an error occurred while loading the content.

Re: [webanalytics] Re: SPSS/SAS/R requirements on analytics job description

Expand Messages
  • Matt Curtis
    They re mostly using these tools to analyze purchase data (customer transaction data). More advanced web analytics teams will also import their clickstream
    Message 1 of 8 , Feb 15, 2013
    • 0 Attachment
      They're mostly using these tools to analyze purchase data (customer
      transaction data). More advanced web analytics teams will also import their
      clickstream data and use SAS / R to make sense of it.

      SPSS isn't built to handle the amount of data that most retailers /
      publishers have - it's more of an academic tool.

      --
      Matt Curtis
      matt.a.curtis@...
      http://www.linkedin.com/in/mattacurtis


      On Fri, Feb 15, 2013 at 9:00 AM, Judah Phillips <judahphillips@...>wrote:

      > **
      >
      >
      > The folks on my current or past teams have used SAS/R/SPSS. Each tool
      > overlaps in functionality; however, I tend to buy SAS consistently.
      >
      > Keep in mind there are SAS programmers and SAS reporters. Some analysts
      > just use/build the reports, where others actually program SAS. Some of the
      > work I've asked my team to do or they've done on their own include customer
      > segmentations (often into quadrants or some lifecycle state), various
      > predictive models (typically logistic regressions), and various x-nomial
      > distributions and hypothesis tests.
      >
      > The data does need to loaded into SAS; however, I don't concur with the
      > person who said that simple measures of dispersion and probability need
      > SAS. Honestly you can use Excel to do the simple basics, like median and
      > mean and even various types of regressions. In fact, if the data set is
      > small, Excel works fine. That said, I don't typically deal with data sets
      > that can be handled (without aggregation) in Excel, so in this case with
      > "big data" we use various tools like SAS.
      >
      > The web analytics tools, then, become more or less mostly dumb pipes that
      > allow us to collect data to model and report in other tools, like SAS.
      > Since the reporting and capabilities in most web analytics tools suck in a
      > harsh Kaushikian way and are mostly inflexible with limited real analysis
      > capabilities like you find in SAS.
      >
      > That said, there are sure a lot of consultants in web analytics who
      > probably disagree with me that web analytics tools are mostly pipes - and
      > that's great and okay - but then again if you look deeply I'd suggest these
      > consultants have no real experience working in brands and even doing
      > analytics, but certainly pontificate about high-level analytics theory well
      > enough to sell consulting services I'd never buy.
      >
      > Judah
      >
      >
      > [Non-text portions of this message have been removed]
      >
      >
      >


      [Non-text portions of this message have been removed]
    • Judah Phillips
      That s a good question, Mark. In terms of growing your career, it all depends on what you want to do and how you want to grow it. For example, a SAS
      Message 2 of 8 , Feb 15, 2013
      • 0 Attachment
        That's a good question, Mark. In terms of growing your career, it all
        depends on what you want to do and how you want to grow it. For example,
        a SAS programmer in pharma at the analyst level is making more than most
        web analytics directors in agencies. Meanwhile, the dude whose never used
        SAS just sold his analytics company and is awaiting his Tesla.

        The reality is the crazy math and head-twisting stats stuff that most
        people don't understand never makes it out of the team at the detailed
        level of complexity. No one, but the analytics team, really cares about
        statistics behind what you are helping them solve for. They only care about
        the solution in real world terms. And most people would likely get annoyed
        and confused if you start talking about the covariance matrix and how
        clever is your fat tailed auto regressive yadda yadda. At the end of the
        day stochastic means random, so we say random because there's not reason to
        look like a smart ass. And in many cases the common charts and graphs, time
        series, and KPI dashboards sans r-squared or any measures of dispersion are
        perfectly appropriate. Occam's Razor and I'm not talking about Avinash's
        blog (Hi AK! :-)

        It sounds like your clients may not be at the level of maturity where they
        are asking or need the data to answer questions that require SAS. Or maybe
        they don't know what to ask for, so suggesting it might be a nice win for
        you. Anyway, the industry needs tool experts so that data is not crap and
        garbage that goes into the models. In fact, it's really hard to get good
        data, so those who can wield the tools get paid but not as much as SAS
        programmers in pharma. :-)


        [Non-text portions of this message have been removed]
      • Matt Curtis
        A phenomenal response. Listen to this man. ... -- -- Matt Curtis matt.a.curtis@gmail.com http://www.linkedin.com/in/mattacurtis [Non-text portions of this
        Message 3 of 8 , Feb 16, 2013
        • 0 Attachment
          A phenomenal response. Listen to this man.

          On Friday, February 15, 2013, Judah Phillips wrote:

          > **
          >
          >
          > That's a good question, Mark. In terms of growing your career, it all
          > depends on what you want to do and how you want to grow it. For example,
          > a SAS programmer in pharma at the analyst level is making more than most
          > web analytics directors in agencies. Meanwhile, the dude whose never used
          > SAS just sold his analytics company and is awaiting his Tesla.
          >
          > The reality is the crazy math and head-twisting stats stuff that most
          > people don't understand never makes it out of the team at the detailed
          > level of complexity. No one, but the analytics team, really cares about
          > statistics behind what you are helping them solve for. They only care about
          > the solution in real world terms. And most people would likely get annoyed
          > and confused if you start talking about the covariance matrix and how
          > clever is your fat tailed auto regressive yadda yadda. At the end of the
          > day stochastic means random, so we say random because there's not reason to
          > look like a smart ass. And in many cases the common charts and graphs, time
          > series, and KPI dashboards sans r-squared or any measures of dispersion are
          > perfectly appropriate. Occam's Razor and I'm not talking about Avinash's
          > blog (Hi AK! :-)
          >
          > It sounds like your clients may not be at the level of maturity where they
          > are asking or need the data to answer questions that require SAS. Or maybe
          > they don't know what to ask for, so suggesting it might be a nice win for
          > you. Anyway, the industry needs tool experts so that data is not crap and
          > garbage that goes into the models. In fact, it's really hard to get good
          > data, so those who can wield the tools get paid but not as much as SAS
          > programmers in pharma. :-)
          >
          > [Non-text portions of this message have been removed]
          >
          >
          >


          --
          --
          Matt Curtis
          matt.a.curtis@...
          http://www.linkedin.com/in/mattacurtis


          [Non-text portions of this message have been removed]
        • Shannon Callan
          If you decide to learn a statistical software, I advocate learning R. It s free.  You can download it and get a manual from O Reilly, learning it on your own
          Message 4 of 8 , Feb 16, 2013
          • 0 Attachment
            If you decide to learn a statistical software, I advocate learning R.

            It's free.  You can download it and get a manual from O'Reilly, learning it on your own with very little investment. To be valuable to the largest number of employers out there, I'd think you'd want to train yourself in the free version of statistical software, not the pricey one.  You can probably transfer your knowledge of R to SAS if you find yourself needing to use SAS.

            Because R is open source, there are all sorts of packages being written for R (like having lots of apps for your iphone).  There is a lot of work being done on text analytics using R.  I'm not sure what SAS offers in text analytics.  But this is an area of growth and innovation and I think the R community is coming up with some interesting packages.

            I've been attending the MeetUps in Los Angeles for R programmers.  It's a great way to find out what people are doing with R and might be a good way for you to find out if you want to learn it.  




            ________________________________
            From: Data Analyst <analyst_data@...>
            To: "webanalytics@yahoogroups.com" <webanalytics@yahoogroups.com>
            Sent: Thursday, February 14, 2013 7:21 AM
            Subject: [webanalytics] SPSS/SAS/R requirements on analytics job description

            Hi All,

            I wanted to get guidance from the practitioners who are at director analytics level and use SPSS/SAS/R. What are the typical analysis you do with these tools.

            I am interested in the following two verticals
            1) Media/Publishing
            2) Retail.

            It just makes me curious that every senior web analytics job description I come across mentions the knowledge of above tools as a requirement. I do not see any of my clients actively using these tools. Some of the clients use these statistical analysis tools but then they do have a statistician on the staff.

            Thank you,
            Mark

            [Non-text portions of this message have been removed]



            ------------------------------------

            ---------------------------------------
            The Web Analytics Forum
            Founded by Eric T. Peterson (www.webanalyticsdemystified.com)
            Moderated by the Web Analytics Association (www.webanalyticsassociation.org)
            Email moderators at: webanalytics-moderators@yahoogroups.comYahoo! Groups Links



            [Non-text portions of this message have been removed]
          Your message has been successfully submitted and would be delivered to recipients shortly.