Skip to main content

Documentation Index

Fetch the complete documentation index at: https://mintlify.com/apursley2012/observatory/llms.txt

Use this file to discover all available pages before exploring further.

​
Case Studies Page

The Case Studies page (/case-studies) is reached via the navigation label “Field Reports”. It renders the CaseStudyReader component, which displays in-depth project post-mortems in a format inspired by scientific field reports and darkroom photo development — a film-negative sidebar for navigation and a wide content panel for the report body.

​
CaseStudyReader Component

CaseStudyReader (components/aurora/CaseStudyReader.js) has two panels:
  • Film-negative sidebar (left) — a narrow panel listing all case study titles as selectable thumbnails styled to resemble photo negatives. Clicking a thumbnail loads that report in the main panel.
  • Darkroom content panel (right) — a wide reading area with high-contrast, low-ambient styling. Report sections are rendered sequentially from the content array.

​
Content Format

Each case study’s content field is an array of strings. Every string is a section that begins with a capitalised keyword followed by a colon and the section body: 
'KEYWORD: Section body text here.'
The component splits on the first : to render the keyword as a styled heading and the remainder as paragraph text. The four standard keywords used in the demo data are:
KeywordPurpose
INITIAL OBSERVATIONSets the scene — the problem that existed before intervention.
HYPOTHESISStates the proposed solution and expected outcome.
METHODOLOGYDescribes the approach taken and key decisions made.
RESULTSQuantifies the outcome with concrete metrics.

​
Case Studies Data

// components/aurora/CaseStudyReader.js
const caseStudies = [
  {
    id: 'cs1',
    title: 'Project Orion: Scaling Telemetry',
    date: 'Stardate 4592.1',
    content: [
      'INITIAL OBSERVATION: The existing monolithic architecture was failing under the load of 10,000+ concurrent data streams.',
      'HYPOTHESIS: Decoupling the ingestion pipeline into micro-services using Go and Kafka would increase throughput by an order of magnitude.',
      'METHODOLOGY: We began a gradual strangler-fig migration. The first phase isolated the most volatile data stream (atmospheric pressure readings).',
      'RESULTS: After a 3-month exposure period, latency dropped from 400ms to 12ms. The system now handles 50,000+ streams with nominal CPU load.',
    ],
  },
  {
    id: 'cs2',
    title: 'Nebula UI: Design System Refactor',
    date: 'Stardate 4610.4',
    content: [
      'INITIAL OBSERVATION: Inconsistent UI patterns across 14 different internal tools were causing severe operator confusion.',
      'HYPOTHESIS: A centralized, highly constrained design system built on React and Tailwind would unify the visual language.',
      'METHODOLOGY: Audited all existing interfaces. Extracted 24 core components. Enforced strict color contrast ratios for low-light environments.',
      'RESULTS: Development velocity increased by 40%. Operator error rates dropped by 15%. The UI now feels like a single, cohesive instrument panel.',
    ],
  },
];

​
Case Studies at a Glance

Problem: Monolithic architecture failing under 10,000+ concurrent data streams.Solution: Strangler-fig migration to Go + Kafka micro-services, beginning with the most volatile stream.Outcome: Latency reduced from 400 ms → 12 ms. System now handles 50,000+ streams at nominal CPU load.
Problem: Inconsistent UI patterns across 14 internal tools causing operator confusion.Solution: Centralised React + Tailwind design system with 24 core components and strict colour contrast ratios.Outcome: Development velocity up 40%. Operator error rates down 15%. Unified instrument-panel aesthetic.

​
Data Shape Reference

FieldTypeDescription
idstringUnique identifier (e.g. 'cs1'). Used as the URL fragment when deep-linking to a specific report.
titlestringReport title shown in the sidebar thumbnail and as the content panel heading.
datestringDisplay date in Stardate format (e.g. 'Stardate 4592.1'). Purely cosmetic — use any string.
contentstring[]Ordered array of section strings. Each string must begin with 'KEYWORD: ' for correct parsing.

​
Adding a New Case Study

1

Open CaseStudyReader.js

The array is at the top of components/aurora/CaseStudyReader.js.
2

Add a new object

Assign a unique id, write a title, pick a date string, and populate content with at least two section strings.
3

Use the standard keywords

Start each section with one of the four keywords (INITIAL OBSERVATION, HYPOTHESIS, METHODOLOGY, RESULTS) for a consistent report structure. You may add custom keywords, but keep them uppercase and follow the KEYWORD: body format.
If a content string does not contain a : character, the component will render the entire string as a body paragraph with no section heading. Always include the keyword prefix.

Build docs developers (and LLMs) love

Get started for freeTalk to us