Presenting at the SAS SUCCESS User Group · •e.g. John and Jane in the sashelp class dataset sound phonically similar according to Soundex proc sql; select a.name as name1, b.name

Presenting at

Why is it that people don’t like to present?

Image from: http://blog.mbabasecamp.com/wp-content/uploads/2012/07/9-giving-a-presentation.jpeg

Potential barriers to presenting

• “What’s in it for me?”

• What do I talk about?

• “I don’t know enough”

• Time / effort to make a presentation

• Nervous about presenting…expectations…

• Uh… how about Matt’s expectations?

• Others…?

Image from http://levelupliving.com/how-common-is-the-fear-of-public-speaking/

Presenting at eSUG J. Prins

• Add/gain value by working using SGF and SUGI papers

• Learn something new at SUCCESS meetings

• Improve your presentation skills

• Contribute to a stronger and more vibrant local SAS user group

Image from: cyberlearningmauritius.org Presenting at eSUG J. Prins

Give a talk based on a SAS Global Forum/SUGI paper!

(Okay, Okay… let me ‘fess up - I can’t take credit for this idea…)

What do I talk about?

Process

• Pick a SAS Global Forum or SUGI paper

• In 5-20 min do one or more of the following:

– Highlight and summarize key points

– Demonstrate code

– Discuss how you used the paper in your work

– Implications in your field

– Expand on the paper

Presenting at eSUG J. Prins

My Favourite Talk from SAS Global Forum 2013

Paper 257-2013: Top 10 Most Powerful Functions for PROC SQL

What’s SQL?

• SQL = Structured Query Language

– Used in all relational DBMS

• PROC SQL

– Increases flexibility in handling data, multiple-table joining

Profiling the 3 functions that I found most useful …

(1) MONOTONIC

(2) SPEDIS & SOUNDEX

(3) MAX

MONOTONIC function

• e.g. Choose SSNs from the 501th line to the 888th line:

proc sql; select * from ssn_data where monotonic() between 501 and 800 ;quit;

Fuzzy matching using SPEDIS

• Asymmetric spelling distance between two words (here singlet detection)

• Simulated dataset of social security nos.

proc sql; select a.ssn1 as x, monotonic(a.ssn1) as x_obs, b.ssn1 as y, monotonic(b.ssn1) as y_obs from ssn_data as a, ssn_data as b where (x gt y) and (spedis(put( x, z11.), put( y, z11.)) le 25) ;quit;

Yefim Gershteyn.'Use of SPEDIS Function in Finding Specific Values'. SAS Users Group International 25. http://www2.sas.com/proceedings/sugi25/25/cc/25p086.pdf

Fuzzy Matching using SPEDIS

• Output:

SOUNDEX: phonic similarity

• e.g. John and Jane in the sashelp class dataset sound phonically similar according to Soundex

proc sql; select a.name as name1, b.name as name2 from sashelp.class as a, sashelp.class as b where soundex(name1) = soundex(name2) and (name1 gt name2) ;quit;

MAX Function

•Question: Is each treatment is effective for patients? •MAX function returns maximum value

•Coding simpler than using a RETAIN statement

and temporary variables at DATA Step

MAX function

• Output:

proc sql; select id, max(treat1) as effect1 'Effect after Treatment 1', max(treat2) as effect2 'Effect after Treatment 2', max(treat3) as effect3 'Effect after Treatment 3' from hospital_data group by id ;quit;

How can I use this paper?

• MONOTONIC: Sub-setting dataset by row numbers after sorting

• SPEDIS: Finding all records describing same events when values affected by typos. Useful before merging datasets by identifier.

• MAX: Identifying lowest cost for most effective treatment