---
version: "alpha"
name: "Predictive Analytics"
description: "Predictive analytics dashboard. Ideal for landing pages, modern websites. AI-ready template."
colors:
  primary: "#1A1A1A"
  secondary: "#4A4A4A"
  tertiary: "#0066FF"
  neutral: "#FFFFFF"
typography:
  h1:
    fontFamily: System UI stack
    fontSize: 2.25rem
    fontWeight: 700
  body-md:
    fontFamily: System UI stack
    fontSize: 1rem
    fontWeight: 400
  label-caps:
    fontFamily: System UI stack
    fontSize: 0.75rem
    fontWeight: 500
---

## Overview

Predictive analytics dashboard. Ideal for landing pages, modern websites. AI-ready template. Prediction has always been the sexy promise of data visualization. But for most of history, we drew trend lines and pretended they were facts. The 2020s changed that — ML models flooded every dashboard with forecasts, and suddenly designers had to reckon with a concept most users find deeply uncomfortable: uncertainty. Confidence intervals aren't new. Statisticians have used them forever. What's new is putting them in front of product managers and executives who want a single number, not a shaded region that says 'maybe somewhere in here.'

The real design challenge isn't technical. It's psychological. Show too much uncertainty and stakeholders lose trust in the model. Show too little and you're lying by omission. The best predictive visualizations thread this needle — they communicate probability as a spectrum, not a guess. Fan charts, gradient fills, ensemble spaghetti plots. Each approach trades precision for legibility differently.

ML democratized prediction. Every SaaS product now ships some forecast widget. But most get it wrong. They bolt a dashed line onto a time series and call it done. The hard work is designing for the space between 'we think' and 'we know' — making that gap feel informative rather than terrifying.

- Density: 8/10 — Dense
- Variance: 4/10 — Moderate
- Motion: 6/10 — Expressive

- **Style:** BI/Analytics
- **Keywords:** Forecast lines, confidence intervals, trend projections, scenario modeling, AI-driven insights, anomaly detection visualization
- **Era:** 2020s Modern
- **Light/Dark:** ✓ Full / ✓ Full

## Colors

- Palette derived from style keywords and era context


## Typography

- **Display / Hero:** System UI stack (-apple-system, sans-serif) — Weight 700, tight tracking, used for headline impact
- **Body:** System UI stack (-apple-system, sans-serif) — Weight 400, 16px/1.6 line-height, max 72ch per line
- **UI Labels / Captions:** System UI stack (-apple-system, sans-serif) — 0.875rem, weight 500, slight letter-spacing
- **Monospace:** JetBrains Mono — Used for code, metadata, and technical values

Scale:
- Hero: clamp(2.5rem, 5vw, 4rem)
- H1: 2.25rem
- H2: 1.5rem
- Body: 1rem / 1.6
- Small: 0.875rem


## Layout

- **Grid:** CSS Grid primary. Max-width containment: 1280px centered with 1.5rem side padding.
- **Spacing rhythm:** Balanced. Base unit: 0.5rem (8px).
- **Section vertical gaps:** clamp(4rem, 8vw, 8rem).
- **Hero layout:** Split-screen (text left, visual right).
- **Feature sections:** Zig-zag alternating text+image rows. No 3-equal-columns.
- **Mobile collapse:** All multi-column layouts collapse below 768px. No horizontal overflow.
- **z-index contract:** base (0) / sticky-nav (100) / overlay (200) / modal (300) / toast (500).


## Elevation & Depth

Forecast line animation on draw, confidence band fade-in, anomaly pulse alert, smoothing function animations

- **Physics:** Spring — stiffness 120, damping 20. Confident, weighted transitions.
- **Entry animations:** Fade + translate-Y (16px → 0) over 480ms ease-out. Staggered cascades for lists: 100ms between items.
- **Hover states:** Scale(1.03) + shadow lift over 200ms.
- **Page transitions:** Fade + slide (300ms).
- **Performance:** Only transform and opacity animated. No layout-triggering properties.


## Shapes

Base corner radius: 8px. See rounded tokens in front matter for the full scale.


## Components

- **Primary Button:** Subtly rounded (0.5rem) shape. Accent color fill. Hover: 8% darken + subtle lift shadow. Active: -1px translate tactile press. Font weight 600. No outer glows.
- **Secondary / Ghost Button:** Outline variant. 1.5px border in muted color. Text in primary color. Hover: subtle background fill.
- **Cards:** Subtly rounded (0.5rem) corners. Surface background. Subtle shadow (0 2px 12px rgba(0,0,0,0.06)). 1px border stroke.
- **Inputs:** Label above input. 1px border stroke. Focus ring: 2px accent color offset 2px. Error text below in semantic red. No floating labels.
- **Navigation:** Primary surface background. Active item: accent color indicator. Font weight 500 when active.
- **Skeletons:** Shimmer animation matching component dimensions. No circular spinners.
- **Empty States:** Icon-based composition with descriptive text and action button.


## Do's and Don'ts

- No emojis in UI — use icon system only (Lucide, Heroicons)
- No pure black (#000000) — use off-black or charcoal variants
- No oversaturated accent colors (saturation cap: 80%)
- No 3-column equal-width feature layouts — use zig-zag or asymmetric grid
- No `h-screen` — use `min-h-[100dvh]`
- No AI copywriting clichés: "Elevate", "Seamless", "Unleash", "Next-Gen"
- No broken external image links — use picsum.photos or inline SVG
- No generic lorem ipsum in demos

- Do Forecast line distinct
- Do Confidence bands visible
- Do Anomalies highlighted
- Do Scenarios switchable
- Do Predictions dated
- Do Accuracy shown


## Use Case

Landing pages, Modern websites

<!-- Source: https://designmd.app/library/predictive-analytics · designmd.app -->
