Human-study dashboard

Live view of rankings collected through the ShapBPT user-study interface, plus offline ratings from the UniTo lab Excel file.

Total rankings
–
Number of (image, method) rank entries.
Participants
–
Unique subject IDs with at least one ranking.
Expertise profiles
–
Source: live web-study (DB)

Rank distributions per method

Approximate version of the grouped bar chart: each row shows how often each method was placed 1st–4th. Colors reflect rank (green = best, red = worst).

1st place
2nd place
3rd place
4th place

Per-trial coverage

How many users have completed each configured example.

Trial #subjects #rankings

User expertise (self-reported)

1 = new to field, 5 = expert. Only users who opted in are counted.

Friedman & Nemenyi statistics

Non-parametric repeated-measures ANOVA (Friedman χ²) over subjects Γ— methods, followed by post-hoc Nemenyi pairwise comparisons.

Test Statistic p-value Interpretation
Friedman – – Waiting for data…

(p < 0.05 indicates at least one method differs significantly.)

Nemenyi p-value matrix: smaller values (e.g., < 0.05) indicate significant differences in average rank between methods. Diagonal = 1.00.