~ / labs/distribution

distribution.

paste a column of numbers — see the shape. histogram + kernel density, summary stats, q-q vs normal, outlier detection (iqr + z-score), and a soft "is this normal-ish?" verdict. all client-side.

presets
count
80
mean
103.554
median
104.03
std dev
14.259
min
66.19
max
139.85
skewness
-0.176
symmetric
excess kurtosis
-0.231
mesokurtic
── histogram + kde · 8 bins
[66.19, 75.4) · 1[75.4, 84.6) · 6[84.6, 93.81) · 14[93.81, 103.02) · 15[103.02, 112.23) · 20[112.23, 121.44) · 16[121.44, 130.64) · 7[130.64, 139.85) · 1μ 103.554med66.19139.85
── q-q plot vs normal
theoreticalobserved
── box plot
139.8566.19q3 114.5med 104q1 92.9
── quantiles
  • min66.19
  • 5%80.706
  • q1 · 25%92.94
  • median · 50%104.03
  • q3 · 75%114.523
  • 95%124.884
  • 99%132.274
  • max139.85
  • iqr21.582
── normality hint
consistent with normal
p ≈ 0.744
  • skewness -0.176 · symmetric
  • excess kurtosis -0.231 · mesokurtic

approximate d'agostino K² test on skewness + kurtosis. for strict inference use a real shapiro-wilk — and treat any "normality test" with healthy scepticism (small samples hide non-normality, big samples flag trivial deviations).

── outliers · 0 by iqr · 0 by |z|>3
no outliers detected