Describe › Standalone plots

Interactive dot diagram (binary cutoff)

An interactive dot diagram lets you drag a cutoff across the data while sensitivity, specificity, PPV, NPV, accuracy and Youden's J update live.

What is Interactive dot diagram (binary cutoff)?

The ROC curve abstracts away the actual data — points become 2D coordinates in (FPR, TPR) space. The interactive dot diagram puts the raw observations back in: positives on top, negatives on bottom, jittered to expose density, with a draggable cutoff line.

Why this matters: choosing a cutoff is a policy decision, not a pure stats decision. You need to see how many false positives and false negatives different cutoffs produce in *this* dataset to make the trade-off honestly.

When should I use Interactive dot diagram (binary cutoff)?

Cutoff exploration after an ROC analysis when you want to see the actual data and the consequences of each cutoff.
Communicating cutoff trade-offs to a non-statistical audience.

What data does it need?

Numeric test value + binary class column (accepts 0/1, pos/neg, yes/no, case/control labels).

What does it report?

Stratified dot plot + vertical cutoff line + live confusion-matrix table with sens / spec / PPV / NPV / accuracy / Youden J + 'Youden optimum' jump-to button.

How do I interpret the result?

The Youden-optimum cutoff is one principled choice; for asymmetric costs (e.g. screening), pick a cutoff that satisfies a fixed-sensitivity or fixed-specificity criterion.