| |
 |
|
|
Science Forum Index » Space - Consult Forum » SAS vs. STATA in terms of automation usability...
Page 1 of 1
|
| Author |
Message |
| jenmoocat... |
 Posted: Thu Jun 19, 2008 5:18 am |
|
|
|
Guest
|
Hi all.
I was tasked by my small company to see which would work best for a
particular use we have, SAS or STATA.
I am a SAS programmer with over 15 years of experience, using it both
as a statistical analysis package and as data manipulation tool.
So..... I am a little biased.
I've read through newsgroups postings about this subject and
understand most of the general comparisons between the two software
packages. However, I still have a question, and it is a little more
specific.
We are looking to create a modularized system of analysis that will
run automatically based on the dataset passed to it. I know how to do
this is SAS, using the macro language, passing parameters hither and
thither.
For example.....
In one case
There is a small dataset with monthly data for one product.
The system would:
Identify that there is only one product.
Idenfity that the data is monthly.
Identify that there are six "spend" variables, based on some naming
convention (prefix of the varname maybe)
Identify that there is no budget constraint variable.
Because there is no budget constraint variable, the system would then
run a module to estimate the budget constraint
It would then run the major module --- estimating aggregate spend for
that one product, in total and then by month.
In a second case:
There is a huge dataset with weekly data for many products.
The system would run as above, only it would recognize that there were
many products and would estimate both an aggregate model across
products and within products. It would also recognize that the data
was weekly and would estimate a total model, plus a monthly model,
plus a weekly model.
To reiterate:
I know how to do all of this in SAS.
I've built automated, macro-driven, modularized systems of data
manipulation and statistical analysis before. And feel pretty
proficient at it.
However, due to the extreme cost of SAS, the company I am now working
for wants to know if STATA might be a better solution.
And while I am confident that it could do the statistical analysis
that we need, I am interested in people's viewpoints about using it
to create an automated system of modules to
1) recognize what a dataset looks like, based on variable naming
conventions (determine how many products, how many different spend-
related variables, the presence or absense of a constraint variable,
frequency of the data)
2) Assign these values to macro variables which would then perform
different analyses (if one product then run only the aggregate
model. If more than one product then run the aggregate model and the
within-product model)
I would really appreciate anyone's advice and experience because I am
not looking forward to growing-pains of learning a totally new syntax
and way to think about datasets -- and then find out that it would be
to klugy for our purpose. Waah.
-jennifer |
|
|
| Back to top |
|
| |
|
Page 1 of 1
All times are GMT - 5 Hours
The time now is Fri Dec 05, 2008 9:17 am
|
|