makino-data-slides

agent
Guvenlik Denetimi
Uyari
Health Uyari
  • License — License: Apache-2.0
  • Description — Repository has a description
  • Active repo — Last push 0 days ago
  • Low visibility — Only 7 GitHub stars
Code Uyari
  • Code scan incomplete — No supported source files were scanned during light audit
Permissions Gecti
  • Permissions — No dangerous permissions requested

Bu listing icin henuz AI raporu yok.

SUMMARY

Data Slides — Turn structured data into investor-grade HTML slide cards. ECharts + static cards, Playwright screenshot.

README.md

makino-data-slides — Data tells the story

数据分析结论直接生成文章级配图,省掉手动做图做排版,带观点带叙事带图表。

License: Apache 2.0
Claude Code
AgentSkills

Features · Install · Usage · Demo · FAQ


Install

cd ~/.claude/skills/
git clone https://github.com/makinotes/makino-data-slides.git

Then type /makino-data-slides in Claude Code.

Usage

/makino-data-slides

Interactive mode: provide your data file + narrative direction, get slide cards.

/makino-data-slides path/to/data.xlsx

Generate slides from a data file (Excel/CSV/JSON/markdown).

Best with Strong Reasoning Models

This skill is not just "draw a chart" — it's analyze data → form opinions → generate visual cards with narrative. The quality of the takeaway line, the choice of comparison angle, the storytelling order — all depend on the model's ability to reason about the data.

Recommended: Use with Claude Code's default model (Opus-class) for best results. Weaker models tend to produce generic takeaways and miss non-obvious patterns in the data.

Features

  • Each card 1080x720, fixed 5-layer structure (header → takeaway → chart → footer)
  • Playwright auto-screenshot, ready to embed in articles
  • Investor-report visual language: hero numbers, stat boxes, ECharts charts, insight callouts
  • Not a chart tool — this is data analysis → opinionated visual cards

What Problem Does This Solve

You finished a data analysis, conclusions scattered across notebooks, and you want article-grade illustrations. Screenshots are ugly, redoing in Figma is slow. This skill turns Excel data directly into publishable visual cards with narrative and opinions — no design tool needed.

When to Use This

Need Use this? Use instead
Single chart (line/bar/pie) No Any charting library
Article → navigable slide deck No /content-to-html
Cover image for article No An image generator
Data analysis → multi-slide cards with narrative Yes

Input

Primary: Excel/CSV with structured data (scores, rankings, time series, category breakdowns).

Also works with JSON, markdown tables, or verbal description of data + conclusions.

Output

Single self-contained HTML file. All CSS and JS inline. Only external dependency: ECharts CDN.

No PowerPoint, no Keynote, no dependencies. Open in any browser, share as a file, or screenshot individual slides.

Each slide is screenshot-ready at 1080x720 via Playwright.

Demo

Demo 1: Logistics — Holiday Order Analysis (4 slides)

Pre/post holiday order volume analysis — grouped bar, area chart, composition breakdown, monthly summary.

Overview Trend
Structure Monthly

Interactive HTML: demo/logistics/demo_slides.html

Demo 2: E-commerce Payment Monthly Report (8 slides)

Full payment operations report from Excel — daily trends, payment methods, category analysis, failure diagnostics, heatmap, user tiers, incident timeline.

Source data: demo/ecommerce-payment/ecommerce_payment_demo.xlsx (8 sheets, 31 days)

Overview Daily Trend
Payment Methods Category Scatter
Failure Analysis Heatmap
User Tiers Incidents

Interactive HTML: demo/ecommerce-payment/ecommerce_payment_slides.html

Design System

5-layer slide structure: header → takeaway → body → footer (fixed 720px height)

Color palette: blue (#1e40af) · red (#dc2626) · green (#059669) · amber (#d97706) · gray (#9ca3af)

Components: hero cards, stat boxes, insight callouts, layer bars, mini tables, finding cards, progress bars

Charts: scatter/bubble, grouped bar, stacked bar, proportion bar (ECharts, animation disabled for screenshot)

Typography: PingFang SC / system font stack, hero numbers 32-40px, body 11-13px

How It Works

Your data  -->  Slide outline (user approves)  -->  HTML generation  -->  Playwright screenshot
  1. Understand: Read data, identify metrics/comparisons/trends
  2. Outline: Propose slide count and content per slide
  3. Generate: Single HTML with ECharts + CSS cards
  4. Review: Open in browser, iterate on spacing/data/readability
  5. Screenshot: Playwright captures each slide at 1080x720

Update

cd ~/.claude/skills/makino-data-slides && git pull

The skill checks for updates automatically on each run.

FAQ

Q: What data formats are supported?
JSON, CSV, markdown tables, or just describe your data in natural language. The skill extracts structure and proposes a slide outline.

Q: Can I customize colors?
The default palette is investor-report style (blue/red/green/amber). You can override per-slide, but the palette is designed to work as a system.

Q: How is this different from a chart tool?
A chart tool draws one chart. /makino-data-slides generates complete slide cards with multiple visual elements, typography, and narrative structure — designed as article illustrations.

Q: How is this different from /content-to-html?
/content-to-html turns a finished article into a presentation or page. /makino-data-slides creates static data cards from raw data — for embedding in articles, not replacing them.

Q: Can I use this for non-Chinese content?
Yes. The typography and layout work with English and Chinese. Just provide your data in any language.

Q: The charts look blank in the screenshot.
ECharts needs a moment to render. The Playwright screenshot script waits 500ms per slide. If charts are complex, increase the wait.

License

Apache 2.0 — see LICENSE and NOTICE

Community & Contact

Bugs, questions and ideas: open an issue. 也可以在公众号「马奇诺」后台留言。

公众号 马奇诺

Author

makino

Yorumlar (0)

Sonuc bulunamadi