googlesheets-charts-graphs - SKILL.md Agent Skill

name: googlesheets-charts-graphs description: Create charts and visualizations in Google Sheets — detect data structure, select chart type, format for clarity, present insights target: google_sheets_agent

Google Sheets: Charts & Graphs

When to Activate

User wants to visualize data, create charts, build dashboards, or add graphs to a spreadsheet.

Step 1: Understand the Data

Before creating any chart, read the data:

GOOGLESHEETS_SEARCH_SPREADSHEETS(query="<name>") → spreadsheet_id
GOOGLESHEETS_GET_SHEET_NAMES(spreadsheetId=spreadsheet_id) → sheets list
GOOGLESHEETS_VALUES_GET(spreadsheetId, range="Sheet1!A1:Z5") → peek at structure

Identify:

Header row (column names)
Data types: numeric, categorical, date/time
Number of columns and rows
What makes sense as X-axis vs Y-axis

Step 2: Select Chart Type

Data Pattern	Best Chart	Why
Categories + values	`BAR` or `COLUMN`	Compare quantities across categories
Time series	`LINE` or `AREA`	Show trends over time
Parts of whole	`PIE`	Show proportions (≤7 categories)
Two numeric variables	`SCATTER`	Show correlation/distribution
Multiple metrics over time	`COMBO`	Lines + bars together
Distribution	`HISTOGRAM`	Show frequency distribution
Hierarchical	`TREEMAP`	Show nested proportions

Rules:

PIE charts: only for ≤7 categories, need exactly 2 columns (labels + values)
LINE charts: best when X-axis is temporal
BAR vs COLUMN: horizontal bars for long category names
SCATTER: only when both axes are numeric

Step 3: Determine Data Range

GOOGLESHEETS_GET_SPREADSHEET_INFO(spreadsheetId) → get sheet_id (numeric, NOT sheet name)

Range rules:

Use A1 notation: 'Sheet1!A1:C20'
First column = domain (labels/categories)
Remaining columns = data series
Include headers: they become legend labels
Single contiguous range only (no comma-separated ranges)

Step 4: Create the Chart

GOOGLESHEETS_CREATE_CHART(
  spreadsheet_id="...",
  sheet_id=0,                    # Numeric sheet ID, NOT name
  chart_type="BAR",
  data_range="Sheet1!A1:C10",
  title="Sales by Region",
  x_axis_title="Region",
  y_axis_title="Revenue ($)",
  legend_position="BOTTOM_LEGEND"
)

Chart type options: BAR, LINE, PIE, COLUMN, AREA, SCATTER, COMBO, STEPPED_AREA, HISTOGRAM, BUBBLE, CANDLESTICK, TREEMAP, WATERFALL, ORG, SCORECARD

Step 5: Present Insights

Don't just create the chart — tell the user what it shows:

Created: "Sales by Region" (Bar Chart)
  Key insights:
  - West region leads with $450K (38% of total)
  - East region grew 22% vs last quarter
  - South region underperforming at $120K

Anti-Patterns

Creating charts without reading data first
Using sheet name instead of numeric sheet_id
PIE chart with 15+ categories (unreadable)
LINE chart for non-temporal categorical data
Comma-separated ranges (not supported)
Creating chart without providing insights