Understanding Statistics Plots
Statistics is easier to interpret when you choose the metric and plot type intentionally.
The same data can look different depending on whether you view counts, proportions, or density.
Choose the metric first
Start by deciding what quantity matches the biological question.
For shared cell detection projects, common options include:
- Total Cells
- ON Cells
% ON- Cells/mm2
- ON Cells/mm2
For independent marker detection projects, use:
- marker-positive cells
- analyzed area
- marker-positive cells/mm2
Do not use % ON, OFF cells, or all-cell density for independent marker detection projects. OFF means rejected candidate, not marker-negative cell.
Heatmaps
Heatmaps are useful for broad overviews across many animals and regions.
They are best for spotting patterns, checking consistency, and finding possible outliers, not for precise reading of small numeric differences.
Bar plots
Bar plots summarize group means for the selected metric.
They are useful when you want a compact comparison across groups and regions. Read them together with variability markers and the underlying animal-level values when possible.
Dot/strip plots
Dot/strip plots show individual animal values.
They are useful for checking whether a group difference is consistent across animals, whether groups overlap, and whether one animal is driving the pattern.
Box/violin plots
Box/violin plots show distribution shape, spread, median, and individual animal values.
They are useful when you care about variability, skew, clustering, or possible outliers rather than only the group mean.
Effect-size plots
Effect-size plots compare each group against a selected baseline using a standardized difference.
They are useful when you want to compare how large a group difference is across regions or metrics, especially when raw values are on different scales.
Scatter plots
Scatter plots compare two selected metrics for the same animal-region measurements.
They are useful for checking relationships, clustering, and outliers, but they should only be used when the two metrics have a meaningful biological or technical relationship.
Mirrored plots
Mirrored plots are useful for direct comparison between two selected metrics, such as counts on one side and percentages or densities on the other.
They work best when both selected metrics are valid for the analysis setup and directly answer the comparison question.
Anatomical representations
Anatomical representations show where a selected metric is concentrated in anatomical space. They use the atlas selected for the project.
Read differences carefully
Before interpreting a strong pattern, ask:
- Is the metric valid for this analysis setup?
- Is the comparison based on the intended subset?
- Does the pattern match what you saw in Results?
A reliable workflow is to pick one biological question, choose the matching metric, choose the plot, and cross-check surprises against Results.