Total New to SAS (Data Minning)

[DigiLana1]

Hello to you ALL,

Just exported my FIRST DATA set in SAS. I have heard about how powerful this Software is, I am excited.

Project 1. Where Should I Begin.

I imported about 75,000 Call Tracker Records looking for Pain Points.. I usually like to begin any Data Analysis with validating..

I am totally New to SAS, how can I be sure the data was imported correctly. What's a good review exercise?

Thanks - Baby Steps  🙂

[TomKari · Accepted solution]

Hi, welcome to SAS!

For anybody involved in analytics, I think this is the best tool out there. And I'm glad you found this community; feel free to use it often, it's a great source of information. Check out the "information for new users" if you haven't already.

You're very wise to double-check your import. My favourite way to do this is to i) export the data I've imported out of SAS, and ii) check that export against my original data using another tool, such as Excel. If your data is voluminous, just take a sample.

Best,
Tom

[DigiLana1]

Thanks Tom for the warm Welcome and useful Insight. That was a great suggestion. I did do a simple Analysis of the data set in Excel as a way to help verify. Also, I did fumble upon the Graph feature in SAS and was able to use the Graph data as a way to validate compared to the initial Excel analysis.

Just like to double check!! 🙂

I am curious to know other ways users Validate their import? Or do they just trust the import process?

I also do not want to over kill...  🙂

[TomKari]

To be honest, I think it's both a science and an art, and every technical tool has its little quirks.

I suggest you start by doing more rather than less checking. Then as time goes by, you'll get a sense of which operations work properly and consistently, and you'll start to ignore them and focus on the types of variables that you use less frequently, or are unusual in other ways.

Always consider the limit variables; the longest, the shortest, the highest, the lowest, etc. Sometimes, when you're using something for the first time, there's no harm in setting up a "synthetic" dataset, with values designed to satisfy you that it's working properly.

Tom