KoliStat · Modern tooling for clinical data

Clinical data analysis,
done in your browser.

XPT, SAS7BDAT, Stata and more — opened, queried, and plotted at DuckDB speed, without a byte leaving your machine.

Table 1 — baseline characteristics by arm · table_one() ▶ live

Characteristic	Placebo	Low Dose	High Dose	Overall
Subjects, N	68	68	68	204
Age, mean (SD)	71.5 (8.1)	72.5 (8.3)	72.5 (7.6)	72.2 (8.0)
Weight (kg), mean (SD)	78.7 (12.7)	75.6 (11.2)	73.9 (10.3)	76.1 (11.6)
Female, n (%)	34 (50.0%)	43 (63.2%)	43 (63.2%)	120 (58.8%)
Male, n (%)	34 (50.0%)	25 (36.8%)	25 (36.8%)	84 (41.2%)

table_one() recomputes this in your browser, with p-values and effect sizes.

Table 2 — bill length on bill depth, one regression per species · lm_fit() ▶ live

Species	Term	Estimate	Std. error	p-value	R²
Adelie	(Intercept)	23.07	3.03	<0.001	0.15
	bill_depth_mm	0.86	0.17	<0.001
Chinstrap	(Intercept)	13.43	5.06	0.010	0.43
	bill_depth_mm	1.92	0.27	<0.001
Gentoo	(Intercept)	17.23	3.28	<0.001	0.41
	bill_depth_mm	2.02	0.22	<0.001

Within each species the slope is positive. Pooled over all three species it is −0.65: Figure 1 shows why.

Figure 1 — bill depth vs bill length, a line per species · VISUALIZE ▶ live

Two layers: DRAW point, then DRAW regression with one fitted line per colour.

The portfolio