Mean Bars¶

Use This When¶

Use mean bars when you want a compact group summary for one or more subject-level summary columns. The plot answers questions such as "is this marker higher in AD than control?" or "which condition has the larger average count?" while still showing the individual subject points when requested.

For a purely descriptive first pass, mean bars are often the fastest plot to read. For a distribution-first view, use distribution plots. For effect-size summaries across many markers, use group comparison plots.

Input Data¶

Mean bars use summary-table data: one row per subject/animal and one numeric column per marker or measurement. A normal imported Batch or Experiment already has this as summary.

Required columns are usually:

Column	Purpose
`AnimalName`	Subject or animal identity used for point means.
`Condition`	Main group assignment.
selected data columns	Numeric measurements to plot.

Raw pandas.DataFrame input is supported by the plotting function when you provide group_col and subject_col, or a full groupList.

Main Functions¶

Function	Registry name	What it draws
`plot_mean_bars`	`mean_bars`	One bar chart per selected summary column, with optional subject points and statistical annotations.

Common Options¶

Option family	Parameters	Use
Column selection	`data_cols`, `data_col_contains`, `data_col_regex`, `data_col_exclude`	Choose the summary columns to plot. See column selection.
Row filtering	`filter_by` / `specificity`, `roi`	Restrict the rows or ROI base before plotting.
Grouping	`split_by` / `factor`, `group_col`, `group_cols`	Plot by the main group list or by one factor in a crossed design. See groups and factors.
Point display	`points`, `point_fill`, `point_edge`, `point_size`, `point_linewidth`	Control whether and how individual observations are overlaid.
Statistics	`comparisons`, `force_nonparametric`, `posthoc`, `posthoc_correction`, `multiple_comparison`, `ns`	Control group tests and annotations.
Saving	`save`, `save_normality`, `normality_dpi`, `dry_run`	Write figures/statistics, save normality checks, or compute statistics without rendering. See saving.

auto_style=True is useful for crossed designs where two groups share a color. It automatically varies fill or hatch styles so conditions remain visually distinct.

Outputs¶

plot_mean_bars returns the shared plotting-run result dictionary. In dry_run=True mode, it returns a pandas.DataFrame of computed statistics without drawing figures.

When save=True, PyFLASH writes SVG bar plots below the object's figure folder, normally in a Bars subfolder. When statistics are run, CSV files are saved next to the bar SVGs. When save_normality=True, normality-check PNGs can also be written next to the corresponding plot.

mean_bars is classified as describe-layer covered: when the report collector is armed, the shared statistics engine emits structured comparison records.

Examples¶

Plot one exact summary column:

from PyFLASH.plotting import plot_mean_bars

result = plot_mean_bars(
    batch,
    data_cols=["GFAP_Count"],
    save=False,
)

print(result.keys())

Find columns by name fragments and save the figures:

plot_mean_bars(
    batch,
    data_col_contains=["GFAP", "Iba1"],
    data_col_exclude="Raw",
    comparisons=["1-2"],
    posthoc="Dunn",
    posthoc_correction="Bonferroni",
)

Use a raw table:

plot_mean_bars(
    df,
    data_cols=["GFAP_Count"],
    group_col="Diagnosis",
    subject_col="Subject ID",
    save=False,
)

Interpretation¶

Bars show group means, not the whole distribution. Use the overlaid points to check group size and spread. A small group with one extreme point can move a mean substantially, so read the points and statistics CSV together.

Statistical annotations are pairwise tests chosen by the shared PyFLASH statistics path. If you need a many-marker result table with effect sizes, multiple-testing control, manifests, and optional superplots, use the group_comparison pipeline.