5 JMP Add-ins for Statistical Modeling
JMP’s visual and interactive environment has long been a favourite for scientists, engineers, and analysts who want to explore data deeply and intuitively. But what happens when your modeling needs become more complex when traditional regression or logistic models are not enough, or when you are dealing with large, imbalanced datasets or too many predictor variables?
That is where JMP’s Marketplace becomes a game changer. A curated platform offering free add-ins developed by JMP, partners, and power users, the Marketplace extends JMP’s capabilities far beyond the base installation. For those engaged in serious statistical modeling, whether in academia, pharma, manufacturing, finance, or clinical domains, a few add-ins stand out not just for their popularity, but for their ability to streamline challenging workflows and bring cutting-edge analytics within easy reach.
Let us dive into five standout add-ins that are shaping how advanced modeling is done in JMP today.
Torch Deep Learning for JMP® Pro
The first is the Torch Deep Learning for JMP Pro add-in. Built for JMP Pro users, it quietly brings the power of PyTorch, one of the world’s leading deep learning frameworks into the visual world of JMP. While many deep learning tools require heavy programming, this add-in provides a structured interface where users can train, validate, and deploy deep learning models with minimal or no coding.
It is particularly useful in contexts where traditional statistical models struggle to capture non-linear relationships or hierarchical patterns in data. For instance, imagine trying to build a predictive model for biological assays using image data or time series signals from wearables. This add-in supports various neural network architectures, handles GPU acceleration, and includes dashboards for training curves and performance monitoring. Researchers in fields like life sciences, material sciences, and advanced manufacturing will find it a natural fit, especially when data richness outpaces what linear models can handle.
XGBoost for JMP® Pro
Next is the XGBoost for JMP Pro add-inconsistently one of the most downloaded machine learning tools on the JMP Marketplace. XGBoost has earned its reputation as the go-to algorithm in Kaggle competitions, and this add-in brings that performance edge into JMP’s modeling ecosystem.
It caters to users who need predictive power across classification and regression tasks think churn prediction in telecom, credit scoring in banking, or even product quality estimation in high-precision manufacturing. With built-in tools for repeated k-fold cross-validation, hyperparameter tuning, confusion matrix analysis, and ROC/AUC outputs, this add-in is remarkably complete. What’s most impressive is that it wraps all of this in a user-friendly wizard, making it accessible to both seasoned data scientists and newer analysts exploring machine learning for the first time. Even without writing a single line of code, users can model complex relationships, validate them rigorously, and deploy them for real-world decisions.
Predictor Selection Assistant
The Predictor Selection Assistant is an often-underrated gem. In many modeling scenarios, users are overwhelmed with tens or hundreds of potential input variables especially in clinical trial analysis, multivariate experiments, or customer analytics. Choosing the right subset of predictors is both an art and a science. This add-in brings structure to that process.
It leverages platforms like Bootstrap Forests and JMP’s Predictor Screening to score variables based on their relative influence, variance explained, and predictive potential. It also allows stepwise selection approaches to be tested and compared visually. For pharmaceutical researchers building QbD models or environmental scientists filtering climate variables, this tool saves time and improves model clarity. Beyond performance, it helps avoid overfitting and multicollinearity, which can erode trust in statistical conclusions. And by surfacing the most meaningful factors, it also sharpens scientific interpretation.
Recommended by LinkedIn
Imbalanced Classification
Handling imbalanced datasets is one of the toughest challenges in classification modeling, and the Imbalanced Classification add-in is designed to meet this head-on. From fraud detection to rare disease identification, analysts often work with datasets where the minority class the one of real interest is dwarfed by the majority.
This add-in integrates several popular resampling techniques such as SMOTE (Synthetic Minority Oversampling Technique), Tomek Links (for noise removal), and random over/under-sampling strategies. These preprocessing tools can be applied before building models in JMP Pro, helping ensure that your models aren’t just accurate overall, but also sensitive and specific to rare but critical events. The add-in is especially useful in pharmacovigilance, insurance fraud analysis, and predictive maintenance, where catching the rare case is more important than broad accuracy. With this tool, users can prepare more balanced training data, improve recall and precision, and explore how different balancing methods affect downstream results.
Calculate Effect Sizes
Finally, the Calculate Effect Sizes add-in answers a common concern among data-savvy professionals: how to go beyond p-values. While a p-value can indicate statistical significance, it doesn't tell you whether the effect is practically meaningful. For researchers submitting to peer-reviewed journals or making regulatory submissions, effect sizes are often mandatory.
This add-in calculates several important metrics: Eta Squared, Partial Eta Squared, and both biased and unbiased Omega Squared. These helps quantify how much variance in the outcome can be attributed to a factor, making your model results more interpretable and defensible. It is especially helpful in educational research, social science experiments, and medical trials where understanding "how much" impact a treatment has is just as important as knowing "whether" it works. It is also a valuable teaching tool many educators use this add-in to help students connect statistical theory to real-world interpretation.
Get them now!
Each of these add-ins reflects JMP’s broader philosophy: making advanced analytics intuitive, accessible, and powerful without sacrificing statistical rigor. Whether you are visualizing complex interactions, optimizing predictive performance, or interpreting model effects in context, these tools empower you to go deeper, faster.
The best part?
All five are free to download and integrate seamlessly into your JMP or JMP Pro environment. They come with sample data, documentation, and an interface that speaks the language of scientific exploration not programming syntax.
If your work involves modeling any kind of structured data whether from a lab, a production line, a clinic, or a customer feedback portal these add-ins are well worth your time. They extend JMP’s native capabilities in smart, user-centric ways, helping you unlock insights that might otherwise remain hidden.
Have you used any of these?
We would love to hear how they have helped in your modeling workflows if you have favorites not listed here.