跳至正文
OpenAgentSkill
浏览已开放 · 目录记录日期 2026-11-21 · 自动更新和账户服务暂时关闭。
设计与 UI

r-visuals

R visual creation and ggplot2 patterns for PBIR reports. Automatically invoke when the user mentions "R visual", "ggplot2", "ggplot in Power BI", or asks to "create an R visual", "add an R chart", "write an R visual script", "inject an R script into Power BI".

data-goblin · 已有 AI 审核记录

安装与使用

运行命令前先阅读来源与依赖说明。复制命令不会自动安装或执行。

npx skills add data-goblin/power-bi-agentic-development --skill r-visuals

技能说明


name: r-visuals description: R visual creation and ggplot2 patterns for PBIR reports. Automatically invoke when the user mentions "R visual", "ggplot2", "ggplot in Power BI", or asks to "create an R visual", "add an R chart", "write an R visual script", "inject an R script into Power BI".

R Visuals in Power BI (PBIR)

Use pbir for every report mutation. Read PBIR metadata only for diagnosis. If pbir is unavailable or lacks an operation, stop and report the gap; never edit report JSON directly.

R visuals execute R scripts (primarily ggplot2) to render static PNG images on the Power BI canvas. ggplot2 is the preferred library -- its grammar of graphics approach produces clean, publication-quality statistical visualizations with less code. R is particularly strong for statistical visualizations.

Visual Identity

  • visualType: scriptVisual
  • Data role: Values (columns and measures, multiple allowed)
  • Data variable: dataset (data.frame, auto-injected)
  • Row limit: 150,000 rows
  • Output: Static PNG at 72 DPI -- no interactivity

Workflow: Creating an R Visual

Step 1: Add the Visual

pbir add visual scriptVisual "Report.Report/Page.Page" --name RevenueByDateR \
  --data "Values:Sales.Date" --data "Values:Sales.Revenue"

Step 2: Write the Script

library(ggplot2)

p <- ggplot(dataset, aes(x=Date, y=Sales)) +
  geom_col(fill="#5B8DBE") +
  theme_minimal(base_size=12) +
  theme(panel.grid.major.x=element_blank())

print(p)  # MANDATORY for ggplot2

Critical rules:

  • print(p) is mandatory for ggplot2 objects -- they do not auto-display in Power BI
  • dataset is auto-injected as a data.frame; do not create it
  • Access columns by index (dataset[,1]) to avoid name escaping issues
  • Use backticks for column names with spaces: dataset$`Order Lines`

Step 2b: Review

Before presenting the script to the user, dispatch the r-reviewer agent to validate correctness and provide design feedback.

Step 3: Inject the Script

pbir visuals r "Report.Report/Page.Page/RevenueByDateR.Visual" --script-file chart.r

The CLI handles PBIR string escaping.

Step 4: Validate

pbir visuals bind "Report.Report/Page.Page/RevenueByDateR.Visual" --show
pbir validate "Report.Report" --all

PBIR Format

For read-only diagnosis, scripts are stored in visual.objects.script[0].properties:

{
  "source": {"expr": {"Literal": {"Value": "'library(ggplot2)\\n...\\nprint(p)'"}}},
  "provider": {"expr": {"Literal": {"Value": "'R'"}}}
}

Identical structure to Python visuals except visualType is scriptVisual and provider is 'R'.

Supported Packages

Power BI Service (R 4.3.3)

PackageVersionPurpose
ggplot23.5.1Grammar of graphics
dplyr1.1.4Data manipulation
tidyr1.3.1Data tidying
ggrepel0.9.5Non-overlapping labels
patchwork1.2.0Compose multiple plots
cowplot1.1.3Publication-quality plots
corrplot0.94Correlation matrices
viridis0.6.5Color scales
RColorBrewer1.1-3Color palettes
forecast8.23.0Time series forecasting
pheatmap1.0.12Heatmaps
treemap2.4-4Treemaps
lattice0.22-6Trellis graphics

~1000 CRAN packages available. Not supported: packages requiring networking (RgoogleMaps, mailR).

Full package list: https://learn.microsoft.com/power-bi/connect-data/service-r-packages-support

Desktop

Any locally installed R package works without restriction. R must be installed separately.

Best Practices

  1. Always call print(p) -- ggplot2 objects require explicit printing
  2. Guard against empty data -- if (nrow(dataset) == 0) { plot.new(); text(0.5, 0.5, "No data") }
  3. Use index-based column access -- dataset[,1] avoids name escaping issues
  4. Use theme_minimal() -- clean aesthetic that works well with Power BI
  5. Factor categorical variables -- control sort order explicitly with factor()
  6. Use hex colors matching the report theme
  7. Set margins -- plot.margin=margin(t, r, b, l) to prevent clipping
  8. Keep scripts concise -- 5-min timeout Desktop, 1-min Service

Limitations

ConstraintDesktopService
OutputStatic PNG, 72 DPIStatic PNG, 72 DPI
Timeout5 minutes1 minute
Row limit150,000150,000
Output size2 MB30 MB
NetworkingUnrestrictedBlocked
GatewayPersonal onlyPersonal only
Cross-filter FROMNot supportedNot supported
Receive cross-filterYesYes
Publish to webNot supportedNot supported
Embed (app-owns-data)Not supportedNot supported

Script Structure Template

library(ggplot2)

# 1. Guard against empty data
if (nrow(dataset) == 0) {
  plot.new()
  text(0.5, 0.5, "No data available", cex=1.5)
} else {
  # 2. Data preparation (index-based access)
  df <- data.frame(
    category = dataset[,1],
    value = dataset[,2]
  )

  # 3. Create visualization
  p <- ggplot(df, aes(x=reorder(category, -value), y=value)) +
    geom_col(fill="#5B8DBE", width=0.7) +
    theme_minimal(base_size=12) +
    theme(
      panel.grid.major.x = element_blank(),
      axis.title = element_blank()
    )

  # 4. Render
  print(p)
}

R vs Python Syntax Reference

For the language-choice decision, see the "When to Use a Script Visual" section above. This table covers only mechanical syntax differences for scripts already committed to R:

AspectR (scriptVisual)Python (pythonVisual)
Render callprint(p)plt.show()
Column accessdataset[,1] or dataset$coldataset.iloc[:,0] or dataset["col"]
Empty guardif (nrow(dataset) == 0)if len(dataset) == 0:
Factor/category orderfactor(x, levels=...)pd.Categorical(x, categories=...)
Runtime
阅读完整来源 ↗