A = Avant-garde Absurdist novel
B = Baby board book
Are you actually advocating Emblem over SAS for GLM?!?!
Just making sure I understood what you are trying to say.
Disclosure: itâs been years since Iâve used either of these tools, but your question, to me, sounds like, âAre you actually advocating a garlic press over a chefâs knife for crushing/mincing garlic?!?!â
Maybe things have changed, but Iâve always viewed Emblem as what Alton Brown would call a âunitaskerâ - itâs a GLM machine. It makes things easy on the user (maybe too easy) and it does its thing pretty well but it doesnât do much anything else. Whereas, SAS is a multitasker where GLM is just one small thing it can do. 
No. Do you disagree that a complete novice could do a GLM faster with Emblem than with SAS?
(FTR: I wouldnât advocate SAS over Emblem, either. A lot depends on what the need is, what resources are available, the overall business data structure, and other related business goals might be.)
Very much this.
Last I heard, the owners of Emblem were looking to expand the range of model types (like using a neural net) than just GLMâs.
And Emblem is limited; some regulators want a particular goodness-of-fit statistic that is not (directly) available within Emblem (unless things have changed). SAS usually has these sorts of things either embedded in the model command (you have to know which switch to âflipâ to have SAS do the calculations; if not provided by default) or thereâs a separate procedure that likely has what you want (but its syntax could be convoluted).
I think that Tableau is on the list because there is no end user adaptation. Developers love it, it is very versatile, and it is visually appealing. I donât know of a single person who receives reports that wants to use it. It might be that my population is biased because almost everyone I know gets reports in excel which they then like to manipulate further (or copy data out of or whatever). So the complaints come from the end users who have to log into one more damn thing to see what they used to get in an emailed Excel file. I donât have much experience other than it was a pain to try to get my data into the tool in the first place. That was more because of the way our servers interact here than the tool. It was also overkill for the one off ad hoc reporting I was doing at the time. I donât need a bunch of pretty graphs and drill downs to show a final value of 6%.
2 Likes
The business world is about networks and connections. The big consulting firms, like Mckinsey, have a huge network of ex employees in all kinds of powerful positions (CEOs, board, politicians, etc.) So despite the fact that they have might have a less then stellar reputation, they are still able to milk large corps and governments.
As for SAS, sometimes itâs good for your employees to use outdated tools because they are less likely to quit the profession and go work for a tech firm that typically uses open source packages in python or R. Imagine you spent years developing young employees and theyâre using the latest open source tech. They become a huge risk for you because they might get hired for more money in another industry that values what theyâve learned. Nobody really seems to use SAS outside of a few industries like insurance and pharma.