WUSS will once again be offering a full menu of optional pre- and post-conference training classes on a variety of topics taught by seasoned experts. Classes are not included in conference registration and must be added separately.
Each half-day class is just $200 with your WUSS 2025 conference registration (or $275 if registering for classes only). This is a terrific value given the quality of instruction for which WUSS has become well-known!
Don’t miss this chance to maximize your experience at WUSS 2025!
| Course Title (click for description) | Instructor(s) (click for bio) |
Time |
| Wednesday, Sep. 3, 2025 – HALF DAY MORNING | ||
| Hands-on Workshop (HOW*) – Mastering the Machine Learning (ML) Toolkit: Training, Tuning, Evaluating & Interpreting Predictive Models with Python | Ryan Paul Lafler | 8:00 AM – 11:30 AM |
| Driving Miss Data: Data-Driven Techniques | Richann Jean Watson | 8:00 AM – 11:30 AM |
| SAS Arrays To Save the Day! | Troy Martin Hughes | 8:00 AM – 11:30 AM |
| Friday, Sep. 5, 2025 – HALF DAY AFTERNOON | ||
| Hands-on Workshop (HOW*) – R You Ready? Applied Statistics with GitHub Copilot | Chary Akmyradov & Lida Gharibvand |
1:30 PM – 5:00 PM |
| Beyond the Basics of SAS Macro | Michele Ensor (SAS) | 1:30 PM – 5:00 PM |
| Bee-yond the Basics: Harnessing SAS, SQL, and Python for Data Analytics in Pharma | Charu Shankar (SAS) | 1:30 PM – 5:00 PM |
Course Descriptions
Hands-on Workshop (HOW) – Mastering the Machine Learning (ML) Toolkit: Training, Tuning, Evaluating & Interpreting Predictive Models with Python
Ryan Paul Lafler
Wednesday, September 3, 2025
8:00 AM – 11:30 AM PDT
This hands-on, half-day workshop is designed for data scientists, statisticians, programmers, ML engineers, and researchers seeking to train and fine-tune supervised Machine Learning models using Python. Attendees will gain practical experience with Python s open-source libraries to build, fine-tune, and evaluate supervised models for classification and regression tailored to real-world predictive needs.
To help attendees master the Machine Learning toolkit, this workshop guides attendees through developing a range of supervised ML models, mitigating overfitting and underfitting, evaluating model performance, and interpreting results and feature significance. Centered around Python s Scikit-Learn (Sklearn) ecosystem, this workshop shows attendees an applied, model-driven approach with key concepts including preparing data for ML models, automating data workflows with Sklearn pipelines, balancing model complexity and interpretability, understanding bias-variance tradeoffs, and comparing statistical models to ML algorithms.
Attendees will learn essential data cleaning techniques, perform Exploratory Data Analysis (EDA) to visualize and understand feature relationships, and build end-to-end Scikit-Learn pipelines. ML Algorithms developed in this workshop include OLS regression, LASSO and Ridge regression, Decision Trees, Random Forests, and Gradient-Boosted Ensembles for classification and regression. Topics include hyperparameter fine-tuning, feature selection and importance, handling model complexity, and strategies for boosting model performance on unseen data.
All attendees will receive PDF slides, an interactive Jupyter Notebook containing the workshop’s documented Python code, and the practical skills to confidently train, optimize, and evaluate predictive models for data-driven AI workflows. Attendees will gain practical experience from using Python libraries including Scikit-Learn, Statsmodels, Pandas, NumPy, SciPy, Matplotlib, and Seaborn.
Driving Miss Data: Data-Driven Techniques
Richann Jean Watson
Wednesday, September 3, 2025
8:00 AM – 11:30 AM PDT
We have all been there. We write a program based on the data we have. Then, we get new data and we must update the program. Making these updates can be time consuming. Not only must you update the production version of the program, but someone must also update any associated validation or QC programs. Wouldn t it be nice if there were ways around this? This is where data-driven techniques come in handy. Using detailed examples, you will learn how to write robust code that is ready to handle an unexpected bend in the road! This half-day course will cover advanced techniques such as: discovering and using information about data sets and variables even if it’s not known in advance; generating dynamic formats that are based on the data instead of hard-coded into your program; using complex looping structures to control your program flow based on the data; building code on the fly, even from within a DATA step; and much more!
SAS Arrays To Save the Day!
Troy Martin Hughes
Wednesday, September 3, 2025
8:00 AM – 11:30 AM PDT
Attend and receive a FREE copy of the author s 550-page book, SAS Data-Driven Development: From Abstract Design to Dynamic Functionality, Second Edition, released in 2022! Students will receive the physical book at the training.
SAS arrays containerize linear, tabular, and multidimensional data into same-type data structures. Arrays can be iterated using various DO loops, and can greatly simplify software by reducing code redundancy and complexity, which in turn maximizes software readability and maintainability. In many cases, an array can be declared and used in lieu of a series of variables, such as when the OF operator is leveraged by built-in functions and subroutines. Moreover, arrays represent the only method to pass multi-element arguments to and from user-defined functions and subroutines created using PROC FCMP (aka, the SAS Function Compiler). This course summarily demonstrates array syntax and advantages in both the DATA step and PROC FCMP.
DATA step array functionality includes:
- declaration and usage (e.g., DO loop) of explicitly indexed arrays
- declaration and usage (e.g., DO OVER loop) of implicitly indexed arrays
- usage of _I_ automatic variable to increment loops
- usage of the DIM, LBOUND, and HBOUND functions to evaluate array dimensions
- declaration of _TEMPORARY_ arrays and their benefits
- use of the OF operator to reference arrays within built-in functions and subroutines
- advanced array syntax to merge data sets and perform lookup operations
PROC FCMP array functionality additionally includes:
- passing a dynamic array to a user-defined function or subroutine
- returning a dynamic array from a user-defined function or subroutine (using the OUTARGS statement)
- declaring a static array inside a user-defined function or subroutine
- usage of the READ_ARRAY function to ingest arrays into FCMP, and the WRITE_ARRAY function to convert arrays into data sets
- usage of the DYNAMIC_ARRAY built-in subroutine (and NOSYMBOLS option) to declare a dynamic array
- conversion between arrays and hash objects
- limitations (and workarounds) of arrays within PROC FCMP
Hands-on Workshop (HOW) – R You Ready? Applied Statistics with GitHub Copilot
Chary Akmyradov, Lida Gharibvand
Friday, September 5, 2025
1:30 PM – 5:00 PM PDT
This hands-on workshop is designed for beginner to intermediate learners seeking to build a strong foundation in statistics while developing practical data analysis skills using R. By the end of this session, participants will be able to approach real-world data problems with statistical confidence and computational efficiency using modern tools like Posit (RStudio) and GitHub Copilot.
The course begins with essential statistical concepts such as types of data, descriptive statistics, visualization, and the importance of the normal distribution, Central Limit Theorem, and sampling distributions. Students will then be guided through the logic of inference covering confidence intervals, hypothesis testing, p-values, and the interpretation of statistical vs. clinical significance.
A unique aspect of this training is the integration of GitHub Copilot, an AI-powered coding assistant, to demonstrate how students can proactively generate, modify, and debug R code in real time. The course will emphasize not only what to code but how to think through the problem first, using Copilot to enhance productivity and reduce the time from idea to insight.
Throughout the session, learners will work with R packages including ggplot2, dplyr, readr, infer, and broom, gaining exposure to tidyverse workflows and reproducible analysis. They will conduct exploratory data analysis (EDA), perform t-tests (independent and paired), ANOVA, and build simple linear regression models. Finally, students will learn to interpret and communicate statistical findings effectively.
Whether you re new to R or looking to modernize your workflow with AI tools, this session will equip you with a robust statistical mindset and a practical coding toolkit to tackle real-world data challenges.
Beyond the Basics of SAS Macro
Michele Ensor (SAS)
Friday, September 5, 2025
1:30 PM – 5:00 PM PDT
Do you use programs containing ampersands and percent signs? If so, you are using the SAS macro facility. It allows you to perform text substitution through the use of macro variables and macro programs. Macros make life easier for the programmer by automating and customizing the generation of SAS code. This tutorial enhances basic understanding of the macro facility by starting with a review of the %LET statement, CALL SYMPUTX routine, INTO clause and macro program syntax, including conditional and iterative processing statements. The tutorial continues with discussion of double periods, multiple ampersands, macro quoting, and the macro IN operator. It will conclude with macro examples of SAS I/O functions, dictionary tables, DOSUBL function, and branching within a macro program.
Bee-yond the Basics: Harnessing SAS, SQL, and Python for Data Analytics in Pharma
Charu Shankar (SAS)
Friday, September 5, 2025
1:30 PM – 5:00 PM PDT
This seminar outlines a structured, five-step approach to data processing Access, Discovery, Manipulation, Analysis, and Reporting applied to bumblebee data as a creative parallel to pharmaceutical analytics.
This approach showcases the complementary strengths of SAS, SQL, and Python in handling diverse analytical tasks. SAS excels in managing large datasets and creating polished reports with its robust data integration and visualization capabilities. SQL demonstrates its power in querying, aggregating, and organizing relational data efficiently, making it indispensable for structured data exploration and filtering. Python shines in its flexibility and scalability, offering advanced analytics, machine learning capabilities, and dynamic visualizations. By leveraging these tools together in SAS Viya Workbench, analysts can create an efficient, end-to-end pipeline tailored for both exploratory and production workflows, bridging the gap between traditional business intelligence and modern data science techniques.
