How to run it 01 Paste the column. Commas, spaces and newlines all separate numbers. In the Auto EDA view, NA, - or ? marks a missing cell, and a first row of names becomes the column headers. 02 Read the cards. The mode card lists every value tied for the highest count, or None; the outlier card lists values outside the 1.5 × IQR fence. 03 Click Auto EDA for the quality view: valid and missing counts per column, the five-number summary, a histogram, skewness and excess kurtosis in words, and the Jarque–Bera note once n ≥ 20. 04 Export or hand off. CSV and Markdown tables download; the vector can be sent to Tests or Regression .
Worked examples Paste 12, 15, 15, 18, 22, 15, 19, 40 into the panel.
Card
Readout
n · mean · median
8 · 19.5 · 16.5
mode · min · max · range
15 · 12 · 40 · 28
sample SD s · population SD σ
8.83176 · 8.26136
SEM · CV
3.1225 · 0.452911
Q1 · Q3 · IQR
15 · 19.75 · 4.75
skewness · excess kurtosis
1.75849 · 1.85847
geometric mean · harmonic mean
18.2283 · 17.317
outliers (IQR)
40
Auto EDA reports the five-number summary 12 | 15 | 16.5 | 19.75 | 40 , one outlier at 12.5% , and the shape line "skewness 1.76 (strongly right-skewed)". At n = 8 the normality card returns no p-value; Jarque–Bera wants 20 points.
Limits to keep in mind
Cards round. Every card prints to six significant digits, so 8.83176 is a display string; export the CSV for more precision.
Outliers are a fence, not a verdict. The flag marks values past Q1 − 1.5 × IQR or Q3 + 1.5 × IQR; confirm the source row before deleting anything.
The means that refuse zero Geometric and harmonic means require strictly positive values. Paste 0, 4, 9 and the arithmetic mean and median still read 4.33333 and 4 , but the geometric and harmonic cards both read NaN — the panel does not drop the zero or fall back to the arithmetic mean. A zero in a growth or rate column is a data question; the cards say so.
Where it helps Screening a pasted measurement log One anomalous reading should not decide the summary: 40 is the only IQR outlier above, and after checking that row the median 16.5 with IQR 4.75 describes the rest of the sample, where the mean 19.5 and SD 8.83176 are pulled by that single point.
Triaging a small spreadsheet Auto EDA keeps missing cells visible: week, yield, temp with one NA shows yield at 7 valid and 1 missing, and the constant temp column is flagged undefined for correlation instead of being reported as r = 0 — a constant carries no co-variation. The surviving pair reads r = 0.9716 , ρ = 1 , cue to fit on Regression .
Privacy The column and its EDA pass compute in your browser; nothing is uploaded and exports are written locally.
References
NIST/SEMATECH, e-Handbook of Statistical Methods , §1.3.5.11 Measures of Skewness and Kurtosis, itl.nist.gov (访问日期:2026-10-01)— skewness and kurtosis conventions.
Wikipedia, Interquartile range , en.wikipedia.org (访问日期:2026-10-01)— the 1.5 × IQR fence behind the outlier card.
Wikipedia, Geometric mean , en.wikipedia.org (访问日期:2026-10-01)— why a zero makes the geometric mean undefined.
Wikipedia, Jarque–Bera test , en.wikipedia.org (访问日期:2026-10-01)— the asymptotic χ²(2) normality check and its small-sample limits.
R Project, quantile: Sample Quantiles , stat.ethz.ch (访问日期:2026-10-01)— the type-7 interpolation behind Q1 and Q3.
Calculators in this hub
Hand-picked tools, one click away. The mini versions compute live and carry your values into the full calculator.
Sources & review
Reviewed by CalcX Editorial Team
Updated 2026-10-01