✨ Darsh

Data analytics, without the ceremony.

---

Extremely simple on the surface. Extremely powerful underneath.

You have data. You want a world-class, responsive dashboard. You probably don't want to spend your evening fighting with HTML, CSS, JavaScript, WebGL configuration, and 400 lines of boilerplate callbacks.

import darsh as da

data = da.load("sales.csv").clean()
da.dashboard(data)

da.get_app().run(port=8501)

No HTML. No CSS. No JavaScript. Zero ceremony.

---

⚡ Installation

pip install darsh

👉 Looking for recipes and guides? Check out HOW_TO_USE.md

---

🧠 The Visual Analytics Mental Model

Just like in Power BI, different charts answer different business questions. The golden rule to remember:

Bar = Compare | Line = Trend | Pie = Share | Scatter = Relationship

Question	Visual Type	Core Purpose	Real-World Example
Which is bigger?	📊 Bar Chart	Compare discrete categories	Sales by Product
Is it changing?	📈 Line Chart	Track continuous trends over time	Monthly Revenue Trajectory
What's my slice?	🥧 Pie / Donut	Parts of a whole (share %)	Market Share by Category
Are they linked?	🔵 Scatter Plot	Discover correlation & clusters	Advertising Spend vs Sales

---

🚀 Feature Highlights

• ⚡ Zero-Boilerplate Web App: Generates a high-performance, responsive web application directly from Python.
• 🌙 Native Dark & Light Themes: Deep space glassmorphic dark mode with interactive instant switcher (or Cmd+D).
• 📱 Streamlit-Style Off-Canvas Sidebar: Smooth collapsible drawer with mobile backdrop blur and keyboard shortcut (Cmd+B).
• 🔄 Real-Time Cross-Filtering: Dropdowns, sliders, and search inputs filter every chart on the canvas simultaneously.
• 🖱️ Interactive Mouse Navigation: Full mouse-pointer scroll zooming (scrollZoom), click-and-drag panning, and double-click zoom reset.
• 📋 Interactive Data Tables: Paginated, searchable tables (da.table(df)).
• 📈 Matplotlib & Plotly Power: Multi-line series, custom markers, line shapes (spline / linear), custom axes, and dimensions (height & width).
• 🌐 3D WebGL Visualization: Hardware-accelerated 3D scatter, line, and bar topology plots with rotational drag.

---

📖 Complete API Reference & Guide

1. Data Ingestion & Auto-Cleaning

# Load local CSV or remote URL
data = da.load("sales.csv")

# Wrap existing DataFrame
data = da.data(df)

# Automatic cleanup (missing values, duplicates, datetimes)
clean_df = da.load("sales.csv").clean().pandas()

---

2. App Setup & Layout Utilities

da.initialize(title="Dashboard", layout="wide", theme="light")

• layout: "centered" (1350px width) or "wide" (100% full-screen canvas).
• theme: "light", "dark", or "auto".

da.page_header(title, subtitle=None, accent_color="#007AFF", badge="🟢 Live")

Adds an Apple-styled top banner with title, subtitle, accent bar, and status badge.

da.header(text) & da.subheader(text)

Hierarchical section headings.

da.row(children)

Responsive horizontal flex grid.

da.table(df, title="Explorer", page_size=8)

Renders an interactive, paginated data table.

---

3. Reactive Sidebar & Cross-Filtering

da.sidebar([
    da.sidebar_header("Global Filters"),
    da.select_filter("Region", column="Region", options=["North America", "Europe", "Asia"]),
    da.select_filter("Product", column="Product", options=["Hardware", "Software", "Services"]),
    da.sidebar_divider(),
    da.range_slider("Max Server Load", column="Server Load (%)", min_val=0, max_val=100, step=5),
    da.search_filter("Search Regions", column="Region", placeholder="Type region name..."),
    da.sidebar_divider(),
    da.sidebar_text("Press Cmd+B to toggle sidebar, Cmd+D for Dark Mode.")
], title="Command Center", width=280, collapsed=False)

---

4. KPI Metric Cards

da.row([
    da.metric("Total Revenue", "$48,250,000", delta="+18.5%"),
    da.metric("Customer Churn", "1.2%", delta="-0.4%"),
    da.metric("Net Promoter Score", "74", delta="+6 pts", delta_color="#5856D6")
])

---

5. 2D Interactive Charts

📈 Line Charts (Multi-Line & Markers)

# Multi-line with list of Y columns
da.line_chart(
    df,
    x="Month",
    y=["Revenue", "Operating Cost", "Net Profit"],
    colors=["#007AFF", "#FF9500", "#34C759"],
    markers=True,
    marker_size=6,
    line_width=3.0,
    title="Financial Performance Trajectory",
    height=400
)

# Grouped multi-line by category
da.line_chart(
    df,
    x="Date",
    y="Active Users",
    color="Product",     # Automatically plots 1 line per Product!
    markers=True,
    title="Active Users by Product Line"
)

📊 Bar Charts (Vertical, Horizontal, Stacked, Clustered)

da.bar_chart(df, x="Region", y="Revenue", color="#007AFF", orientation="v")
da.stacked_bar_chart(df, x="Quarter", y=["Hardware", "Software"], colors=["#007AFF", "#34C759"])

🥧 Pie & Donut Charts

da.donut_chart(df, labels="Category", values="Sales", hole=0.45, colors=["#007AFF", "#5856D6", "#FF9500"])

🔵 Scatter & Bubble Plots

da.scatter_chart(df, x="AdSpend", y="Sales", color="Region", size="CSAT", marker_size=10)

---

6. 3D WebGL Charting Engine

da.scatter_3d(df, x="Users", y="Load", z="CSAT", color="Product", title="3D System Telemetry")
da.line_3d(df, x="Date", y="Revenue", z="Users", color="#007AFF", title="Growth Trajectory 3D")
da.bar_3d(df, x="Region", y="Product", z="Revenue", colorscale="Blues", title="Revenue Matrix 3D")

---

7. Automatic Magic Dashboard

data = da.load("sales.csv").clean()
da.dashboard(data)
da.get_app().run(port=8501)

---

👨‍💻 Creator & Maintainer

Darsh was created by Satyam Rana, a Full-Stack Engineer, AI Practitioner, and Educator passionate about crafting high-performance developer tools and beautiful data systems.

---

📄 License

Darsh is open-source software licensed under the MIT License.