Chart gallery¶

Lines and trends¶

Series followed along an axis, with their bands and right-hand scales.

Lines with an error band¶

df has condition, time (h), signal (a.u.) and se (a.u.). The band spans one se (a.u.) either side of each point.

chart = i.line(df, x='time (h)', y='signal (a.u.)', color='condition',
               error_y='se (a.u.)', markers=True)

Names at the line ends¶

df is the growth table from the first example.

chart = i.line(df, x='time (h)', y='signal (a.u.)', color='condition',
               legend='direct')

Lines on a right-hand axis¶

df is the generation table from the stacked-area example. secondary_y= puts the named series on a right-hand axis with its own scale.

chart = i.line(df, x='year', y='generation (TWh)', color='source',
               secondary_y=['solar'], markers=True)

Stacked areas¶

df has year, source and generation (TWh). Areas stack unless stacked=False.

chart = i.area(df, x='year', y='generation (TWh)', color='source')

Slope chart¶

df has year, country and support (%): two time points per country. format= writes the values as percentages.

chart = i.slope(df, x='year', y='support (%)', group='country', format='{:.0f}%')

Relationships¶

Points against points, with fitted lines.

Scatter with groups¶

df has strain, dose (mg/kg) and response (%).

chart = i.scatter(df, x='dose (mg/kg)', y='response (%)', color='strain')

Scatter coloured by a number¶

df has temperature (°C), oxygen (mg/L) and chlorophyll (µg/L). A numeric color column with many values is drawn as a ramp.

chart = i.scatter(df, x='temperature (°C)', y='oxygen (mg/L)',
                  color='chlorophyll (µg/L)')

Regression with a confidence band¶

df has concentration (µM) and absorbance (AU), two readings per standard.

chart = i.regression(df, x='concentration (µM)', y='absorbance (AU)')

Regression with its equation¶

df is the calibration table from the regression example. equation=True writes the fitted line on the plot.

chart = i.regression(df, x='concentration (µM)', y='absorbance (AU)', equation=True)

Distributions¶

The spread of values, as histograms, curves and dots.

Overlaid histograms¶

df has sex and length (mm), one row per specimen.

chart = i.hist(df, x='length (mm)', color='sex', bins=20)

Filled density curves¶

df has genotype and expression (log2), 80 samples per genotype.

chart = i.kde(df, x='expression (log2)', color='genotype', fill=True)

Cumulative distributions¶

df has group and latency (ms), 50 animals per group.

chart = i.ecdf(df, x='latency (ms)', color='group')

Box plots with points¶

df has treatment and yield (t/ha), 15 plots per treatment.

chart = i.boxplot(df, x='treatment', y='yield (t/ha)', points=True)

Violins¶

df has cohort and grip force (N), 40 people per cohort.

chart = i.violin(df, x='cohort', y='grip force (N)')

Strip plot¶

df has site and soil pH.

chart = i.strip(df, x='site', y='soil pH')

Categories and shares¶

Bars, dots and wedges for counts and totals.

Bars counting rows¶

df has one row per specimen in a species column. Without y, a bar counts its rows.

chart = i.bar(df, x='species')

Grouped bars¶

df has site, season and biomass (g/m2), one row per site and season.

chart = i.bar(df, x='site', y='biomass (g/m2)', color='season')

Mean bars with error bars and points¶

df has group and response (a.u.), one row per animal. agg= takes the mean of each group, and error_y= takes its standard error.

chart = i.bar(df, x='group', y='response (a.u.)', agg='mean',
              error_y='sem', points=True)

Grouped mean bars with error bars and points¶

df has group, sex and response (a.u.), one row per animal. color= puts the bars side by side, and agg=, error_y= and points= work as they do for a single series.

chart = i.bar(df, x='group', y='response (a.u.)', color='sex', agg='mean',
              error_y='sem', points=True)

Lollipop chart¶

df has pathway and score, one row per pathway. orient='h' lays the stems across the page.

chart = i.lollipop(df, x='pathway', y='score', orient='h')

Dumbbell chart¶

df has site, before (%) and after (%), one row per site. x= names the two value columns.

chart = i.dumbbell(df, y='site', x=['before (%)', 'after (%)'])

Waterfall chart¶

df has step and change (k$). The two totals= steps stand from zero; a missing change on a total shows the running total.

chart = i.waterfall(df, x='step', y='change (k$)', totals=['Opening', 'Closing'])

Pie chart¶

df has cell type and cells (n), one row per type. The slices are sized by values=, labelled with their shares, and keyed by names=.

chart = i.pie(df, names='cell type', values='cells (n)')

Donut chart¶

df is the cell-type table from the pie chart. hole=0.5 leaves a hole half the radius of the disc.

chart = i.pie(df, names='cell type', values='cells (n)', hole=0.5, legend='bottom')

Matrices¶

Values in a grid of coloured cells.

Heatmap with a diverging colour bar¶

df is a long table with gene, condition and log2 fold change. A diverging palette= with center=0 puts zero at the middle of the colour bar.

chart = i.heatmap(df, x='condition', y='gene', z='log2 fold change',
                  palette='rdbu', center=0)

Studies and survival¶

Statistics from experiments, trials and cohorts.

Volcano plot¶

df has one row per gene: log2 fold change, a raw p-value and an adjusted fdr. q= classes each point by its adjusted p-value, and highlight= names the genes to label.

chart = i.volcano(df, x='log2 fold change', y='p-value', label='gene', q='fdr',
                  highlight=['Il6', 'Cxcl10'])

Forest plot¶

df has study, odds ratio, lower and upper, one row per study. summary= draws a flagged row as a diamond, and right= lists the text beside the plot: here the estimate with its interval.

chart = i.quick.forest(df, label='study', estimate='odds ratio', lower='lower',
                       upper='upper', summary='pooled', log=True, measure='OR',
                       right=['ci'])

Survival curves¶

df has arm, months and died. died is 1 where the patient died and 0 where follow-up ended first, which is a censored observation.

chart = i.survival(df, time='months', event='died', color='arm')

Panels¶

Several charts in one figure, with shared axes or lettered panels.

Facet panels¶

df has cell line, drug, time (h) and viability (%). facet_col= makes one panel per value of a column.

chart = i.line(df, x='time (h)', y='viability (%)', color='drug',
               facet_col='cell line')

Combined panels¶

df is the growth table, and outcomes is the response table. | places panels side by side and / stacks them.

trace = i.line(df, x='time (h)', y='signal (a.u.)', color='condition',
               error_y='se (a.u.)')
spread = i.boxplot(outcomes, x='group', y='response (a.u.)', points=True)
chart = (trace | spread) / i.hist(outcomes, x='response (a.u.)', color='group', bins=14)