MCDM Scatter Chart 2D/3D

The MCDM Scatter Chart widget plots trials on two chosen axes — 2D and 3D variants — and colors each point by its MCDM rank. It projects a ranking computed from many criteria back onto a familiar scatter of objectives (or method scores), so you can see where the top-ranked designs sit in the trade-off space, not just their order. It is the spatial companion to the MCDM Ranking chart, which lists the same ranking as bars.

The ranking itself is produced by the MCDM methods (TOPSIS, VIKOR, PROMETHEE, and the entropy weighting option); this page covers the widget. For the theory and how to configure weights, see MCDM and the usage guide.

Workflow

1. In the controls, choose the MCDM method, set the criterion weights and the Top-N highlight count, then press Run to compute the ranking.
2. Choose the X and Y axes (and Z in the 3D variant). The options are each objective, plus the selected method's score axes:
   • TOPSIS — TOPSIS Score
   • VIKOR — Q / S / R values
   • PROMETHEE — Phi+ (positive flow), Phi- (negative flow), Phi Net
3. Read the colored ranking. Click any point to open the trial-detail modal, which shows that trial's MCDM rank and score (and, for VIKOR, whether it belongs to the compromise set).

Reading the chart

• Color encodes rank within Top-N. Rank 1 (best) gets the brightest end of the colormap and fades down to rank N; trials outside Top-N are drawn gray as Others. Raise Top-N to color more of the field, lower it to spotlight only the leaders.
• Infeasible trials on a constrained study are drawn in a separate color and can be toggled from the legend.
• The compromise set matters for VIKOR. Solutions satisfying Opricovic & Tzeng's acceptance conditions (C1/C2) are listed by trial number under the chart and flagged in the detail modal — these, not merely "rank 1," are VIKOR's recommended answer.
• Cross-filtering: while a selection filter is active (for example a Parallel Coordinates brush), unselected points are dimmed. Scores are still computed over the full Pareto front, so brushing changes only visual emphasis, never the ranking.

Choosing axes

• Objective vs. objective places the ranking on the usual trade-off plot — the quickest way to confirm the top designs sit where you expect on the Pareto front.
• Score vs. objective (e.g. TOPSIS Score against an objective) shows how the aggregate score trades against a single criterion.
• Score vs. score (e.g. VIKOR S against R) exposes the internal structure of the method — S is the group utility, R the worst individual regret.

Caveats

• The ranking depends entirely on the weights. The colors move when you re-run with different weights; treat the chart as one weighting scenario, not an absolute ordering. Compare a few weightings before committing.
• Only trials with finite values on both (all three) chosen axes are plotted.
• For an exact side-by-side of a shortlist of candidates, pin them and use Comparison Table and Radar Comparison; the scatter is for seeing the whole field at once.