JSerek

quant-correlation

Compute a correlation matrix for two or more numeric/ordinal variables. Primary method: Spearman correlation with permutation p-values and bootstrap CIs. Alternatives: Pearson (continuous + linear), Kendall's tau-b (ordinal/ties-heavy), Distance correlation (non-linear). Produces interactive heatmap, scatter matrix, and plain-language interpretation.

Web & Frontend Listed

quant-design

Project setup and study design for quantitative UX/CX/Market Research. Elicits research questions, hypotheses, and survey structure from the user. Creates analysis_journal.md in the project root — the persistent, human-readable memory file shared with quant-orchestrator across all future analysis sessions. Run ONCE at project start. Re-invoke to add hypotheses or update study design.

quant-model-continuous

OLS regression for continuous DVs with HC3 heteroscedasticity-robust standard errors. Assumption checks (normality, homoscedasticity, Cook's D, VIF), Cohen's f², diagnostic grid, coefficient forest plot, and partial regression plots. Dual-audience HTML report.

quant-model-ordinal

Cumulative Link Model (proportional odds / ordered logit) for ordinal DVs such as Likert 1–5 scales. Proportional odds assumption test, Nagelkerke R², OR forest plot, and predicted probability curves. Dual-audience HTML report.

quant-orchestrator

Post-analysis routing and interpretation for quantitative UX/CX/Market Research. Reads analysis_journal.md and output/data/results.json to detect new results, interpret them in the context of hypotheses, update the journal, and recommend the next analysis step with verbose settings (skill name + per-question answers). Run after every analysis skill completes, or to check project status.

Web & Frontend Listed

quant-scale-reliability

Psychometric reliability analysis for Likert-scale survey batteries. Computes Cronbach's α (with bootstrap CI), McDonald's ω_total, item-total correlations, α-if-item-deleted, average inter-item r, and optional subscale reliability. Flags problematic items and outputs an interactive dual-tab HTML report with three diagnostic plots.

quant-data-prep

Data preparation skill for quantitative UX/CX/Market Research analysis. Use this skill to load, audit, clean, recode, and label survey or research datasets before any analysis. Always run this skill first — even on "pre-cleaned" data. It performs a full data quality audit and surfaces all issues via AskUserQuestion before making any changes. Triggers: "prepare data", "clean data", "load data", "recode", "data prep", "data cleaning", "remove straightliners", "handle missing values", "label variables", "before analysis", "data quality check", "audit data".

Web & Frontend Listed

quant-exploration

Exploratory data analysis (EDA) skill for quantitative UX/CX/Market Research. Produces descriptive statistics, frequency tables, distribution visualizations (histograms, violin/box plots with jitter), and a Spearman correlation heatmap. All outputs are interactive HTML (Plotly) plus PNG exports. Run this skill after quant-data-prep to familiarize the user with the data before committing to any analytical method. Triggers: "explore data", "EDA", "descriptive statistics", "visualize data", "frequency tables", "distribution plots", "correlation heatmap", "data overview", "summary statistics", "box plot", "histogram", "violin plot".

quant-model-multinomial

Multinomial logistic regression for unordered categorical DVs with 3+ categories.

quant-model-nominal

Binary logistic regression with HC3 robust standard errors for binary DVs (0/1, yes/no, aware/unaware).

quant-multi-group-dep

Compare an outcome across three or more dependent (within-subject) measurements — e.g., three timepoints, three conditions rated by the same respondents, or three brand dimensions evaluated within-person. Uses a permutation Friedman test as the omnibus, followed by MANDATORY pairwise post-hoc sign-flip permutation tests with BH FDR correction and Cohen's dz for each pair. Triggers: "three timepoints", "three conditions", "repeated measures", "within-subject", "same respondents across all", "before during after", "track over time", "multiple waves", "repeated measurements", "friedman test", "3+ paired conditions", "longitudinal comparison".

quant-multi-group-ind

Compare a continuous or ordinal variable across three or more independent groups. Uses a permutation Welch F-test (robust to unequal variances) as the omnibus test, followed by MANDATORY pairwise post-hoc comparisons with Benjamini-Hochberg FDR correction and Hedges' g for each pair. Effect sizes are reported for every comparison. Triggers: "compare three groups", "three or more groups", "multiple groups", "one-way ANOVA", "Kruskal-Wallis", "compare markets", "compare brands", "compare segments", "multi-group comparison", "which groups differ", "group differences across categories".

quant-two-group-dep

Compare two dependent (paired) measurements — e.g., pre vs post, before vs after, time 1 vs time 2, or two conditions for the same respondent. Uses a permutation paired t-test (sign-flip test) as the primary method. Reports Cohen's dz with bootstrap CI, individual trajectory plots, and plain-language interpretation. Triggers: "pre post", "before after", "paired comparison", "repeated measurement", "same respondents measured twice", "change over time", "within-subject difference", "two time points", "paired t-test", "dependent samples", "matched samples".

quant-two-group-ind

Compare a continuous or ordinal variable between two independent groups. Uses a permutation Welch t-test (robust to unequal variances and non-normality) as the primary method. Reports effect size (Hedges' g) with bootstrap CI, plus full assumption checks, visualizations, and plain-language interpretation. Triggers: "compare two groups", "difference between groups", "independent samples", "group comparison", "A/B comparison", "male vs female", "test vs control", "two-group t-test", "between-group difference", "is there a difference between".