---

name: basic-charting description: "Build core Chartify charts from tidy pandas DataFrames, choose axis types and plot methods, inspect plotted data, and save or show chart outputs." disable-model-invocation: true metadata: disco-role: operating license: Apache 2.0

Basic Charting

Use this sub-skill when a task asks for core Chartify chart construction from pandas DataFrames: line, scatter, text, area, bar, stacked bar, boxplot, interval, lollipop, parallel, heatmap, density, radar, or second-y-axis charts. This sub-skill owns Chart and RadarChart creation, axis-type selection, plot-method routing, data-shape preparation, labels/legends enough for basic chart workflows, Chart.data inspection, and show/save output calls.

Route palette creation, YAML options/style configuration, callout-heavy annotation, and advanced axis formatting to sibling styling-annotations. Keep maintainer docs, package release, and documentation-build workflows out of this sub-skill.

Read/Run These Files

• Read references/api-reference.md when choosing Chart(...) or RadarChart(...) arguments, selecting a plot method, or checking method signatures and return objects.
• Read references/data-formats.md when converting pandas results into Chartify's tidy input format or debugging grouped/pivoted data.
• Read references/workflows.md when you need compact, self-contained chart construction patterns for common chart families, save/show, data inspection, radar charts, or second-y-axis charts.
• Read references/troubleshooting.md when a plot method is unavailable, dtypes do not match axis types, grouped categorical data orders incorrectly, second-y-axis plotting fails, or PNG/SVG export needs a browser driver.
• Run scripts/chartify_smoke_examples.py to exercise safe tiny examples without notebook display or network access; use --list, --case, and optional --save-html DIR.

Fast Routing Rules

1. Start with import chartify and a pandas DataFrame with named columns.
2. Choose axis types before choosing the plot method:
   • numeric/datetime x with numeric y: Chart(x_axis_type="linear"|"log"|"datetime", y_axis_type="linear"|"log") and line, scatter, text, or area.
   • categorical plus numeric: make the categorical axis explicit with x_axis_type="categorical" for vertical charts or y_axis_type="categorical" for horizontal charts, then use mixed-type methods such as bar, bar_stacked, boxplot, interval, lollipop, parallel, scatter, or text.
   • categorical/categorical heatmap: set both axis types to "categorical".
   • density charts: set one or both axes to "density" for histogram, kde, or hexbin.
   • radar: use RadarChart(...), not Chart(...).
3. Convert groupby/pivot outputs to tidy DataFrames before plotting: use .reset_index() after groupby and pd.melt or stack().reset_index(...) for wide tables.
4. Add basic labels with set_title, set_subtitle, set_source_label, and legend placement with set_legend_location after plotting grouped/color data.
5. Inspect ch.data for a list of plotted Bokeh ColumnDataSource.data dictionaries before saving or returning a chart object.