Individual Value Plot

Anatomy of an Individual Value Plot

An individual value plot shows every single observation as its own point. Unlike a boxplot — which collapses the distribution into five summary numbers — each measurement remains visible. This makes the plot especially useful for small to medium samples where a boxplot has too few points to summarize meaningfully.

Point: Each point represents one measurement. Its vertical position is the value on the Y axis; its horizontal position assigns it to a group (category).

Jitter (horizontal scatter): When several observations share the same Y value, they would otherwise overlap. A small random horizontal offset ("jitter") separates the points visually without changing their Y value. Jitter is purely a readability device and carries no information.

Mean diamond: One diamond per group marks the arithmetic mean. It enables quick comparison of location across groups.

Median tick (optional): A short horizontal tick per group marks the median. Combined with the mean diamond it makes skewness visible at a glance: a clear gap between mean and median indicates an asymmetric distribution.

Connecting line through means (optional): A dashed line connects the group means, surfacing trends across ordered groups (e.g. shift, day, dose level).

Overall mean (optional): A horizontal reference line marks the overall mean across all groups. It is useful for telling at a glance which groups sit above and which sit below the overall level.

In short: points = raw data, diamond = mean, optional tick = median, optional lines = comparison aids. The individual value plot retains more detail than a boxplot and is the right choice when you want to see every single data point.

Methodology

The individual value plot is a descriptive tool. It takes one numeric column (Y) and assigns each observation to a category (X) — either via up to three grouping columns or by comparing several numeric columns side by side.

When should I use it?

• Sample size below 50 per group: boxplots become unreliable with few points; the individual value plot shows every value.
• Discrete or rounded data: clusters on a handful of values are invisible in a boxplot but obvious here.
• Comparing several groups while focusing on individual outliers.
• As a sanity check before formal tests (t-test, ANOVA) to visually inspect the assumptions.

Input Modes

• Multiple columns: Each selected numeric column produces its own scatter of points. Ideal for direct comparison of different measurements.
• Grouped: A single value column is split by a grouping column (e.g. machine, shift, supplier).
• Nested (up to 3 levels): G1, G2 and G3 act combinatorially — every unique combination of group values produces its own column of points (e.g. G1 = shift, G2 = machine → one column per shift/machine combination). The column label lists the group values separated by " | ".

Mean: Arithmetic mean of all values in a group. Sensitive to outliers.

Median: 50 % point of the sorted values. Robust against outliers.

Standard deviation: Shown in the tooltip (n − 1 denominator). A measure of within-group spread.

Practical Example

1. Open the Individual Value Plot module.
2. Select value column = "ResponseTime_ms" and grouping column = "Week".
3. Options: enable the mean diamond and the connecting line through means.
4. Each week appears as a vertical cloud of points; the dashed line traces the trend of the means.
5. W16 shows a visible upward shift with two distinctly higher individual values.

Interpretation

What to look for

• Cloud location: where does the bulk of a group sit relative to the others?
• Within-group spread: a wide vertical cloud means high variability, a tight cloud means low variability.
• Clusters: multiple points at the same Y value (visible through jitter) suggest rounded or discrete measurements.
• Gaps: an empty band inside a group's range can indicate bimodal data or a missing measurement level.
• Isolated points: potential outliers — investigate the root cause, do not delete automatically.
• Mean vs. median: a large gap between the two indicates a skewed distribution.

Pitfalls

Examples

This module ships with the following example datasets — load any of them in the app with a single click.