tag:blogger.com,1999:blog-2956099714546919205.post7733076406385857307..comments2024-03-25T00:59:47.919-07:00Comments on Musings of a forgetful functor: Survival skills for today's analystaxiomOfChoicehttp://www.blogger.com/profile/10274657326813344423noreply@blogger.comBlogger3125tag:blogger.com,1999:blog-2956099714546919205.post-34943067177090211062012-02-24T02:08:56.362-08:002012-02-24T02:08:56.362-08:00I must have missed this post of yours. I share eve...I must have missed this post of yours. I share every thoughts you have here. I have worked in a research institue for 3 years and am working for a GOVT ministry now. I use R on the daily base. Guess what? The work I have done have ALWAYS stand out because of R. For example, quickness to process complicated data analysis and superior visualization (see some examples from my blog http://nzprimarysectortrade.wordpress.com/). They are the value added service but not a simple line or bar plot produced using M$ Excel.<br /><br />I must say that there may not be rooms for analysts without programming skills in 5 years time when any kinds of data are beyond the handle of M$ Excel. Invest in R is a wise move.Weihttps://www.blogger.com/profile/13917273444008448741noreply@blogger.comtag:blogger.com,1999:blog-2956099714546919205.post-77017968521241255532011-07-09T01:07:41.250-07:002011-07-09T01:07:41.250-07:00Al. Thanks very much your comment. I can see that ...Al. Thanks very much your comment. I can see that there are strong feelings from both the Excel and Programmer communities. As for the problems associated with R, I understand completely where you are coming from. In particular, I can understand the frustration that you, as an experienced programmer, face in working with the poor R code of the non-programmer. <br /><br />However, I think that although far from perfect R may be a stepping stone in the history of programming languages. After a time, programming skills may developed in the wider analytic community and languages like Python will become common tools for analysts. Of course, that seems like a highly optimistic view when you spend your days debugging nightmare code created by business majors.<br /><br />Thanks again for your comments.axiomOfChoicehttps://www.blogger.com/profile/10274657326813344423noreply@blogger.comtag:blogger.com,1999:blog-2956099714546919205.post-85525456327950733702011-07-07T21:51:53.255-07:002011-07-07T21:51:53.255-07:00I agree 50%. Excel is overused in my organization...I agree 50%. Excel is overused in my organization too. Some spreadsheets are incredibly complicated, so that it is near impossible to debug them--a few lines of Python or R would be much easier to maintain. Others try to use excel when a database should be used (V/HLOOKUP ARRGH!). <br /><br />If "R seems strange and frightening to the non-programmer", I assure that you that it seems even stranger to programmers! R is one of the nicest statistical DSLs, but it is also an ugly, unpredictable beast that tends to encourage poor coding habits. Code often switches from OO to procedural to functional paradigms in a single script, it has inconsistent naming among packages (and in the R Core too). To get many things done, one often has to inline C code. R is a mess, its packages are more of a mess and most R-code that I see in companies is barf-worthy, bug prone, ad hoc and slow as molasses (having been written by non-programmer Excel converts). I don't agree that we should be convincing non-programmers to try R, they just make a mess. These folks should be encouraged to learn a good SQL database--that would take care of much of the Excel problem. Then, they need a good data analysis platform like SPSS or Stata that will allow them to do proper analyses and make graphs, but won't give them enough rope to hang themselves. <br /><br />Face it, most people just aren't up to learning to PROPERLY program. That doesn't mean they shouldn't use better tools, but I don't think R is the right choice for those who aren't motivated to learn to program. R is hard, and it takes at least a few months for an experienced programmer to learn to use it properly. Most folks are not going to invest the time to learn the basics. You can look forward to trying to debug thousands of lines of R spaghetti code in the coming year if you stay on your soapbox.Anonymousnoreply@blogger.com