Phase 2 Strategy: URL Standardization & Topics Taxonomy

Date: 2025-11-22 Status: Planning & Design (No Code Yet) Goal: Agree on end-state before implementation

Current State Analysis

1. Blog URL Patterns (PROBLEM: 3 Different Patterns)

Current URLs by Series:

Series	URL Pattern	Example	Count
Website Building	`/posts/{slug}/`	`/posts/getting-started-with-github-pages/`	4 posts
Sensor Fleet	`/temp-sensor-{n}/`	`/temp-sensor-00/`, `/temp-sensor-01/`	7 posts
Individual Posts	`/{slug}/` or `/posts/{slug}/`	`/print3d-00/`, `/stock-ticker-00/`	4 posts
Config Default	`/:categories/:title/`	(not used consistently)	N/A

Why This Is A Problem:

❌ Unpredictable: Users can’t guess URLs
❌ Inconsistent branding: Some have /posts/, some don’t, some have numeric codes
❌ SEO dilution: Not grouped under a recognizable path
❌ Maintenance burden: Hard to implement site-wide features (breadcrumbs, series navigation, etc.)
❌ Migration risk: If you want to change patterns later, you need redirects for 3 different formats

2. Categories (Currently Used)

Category	Count	Posts
tutorial	5	Website series (4) + Mermaid (1)
data science	7	Sensor Fleet series (7)
cloudML	1	AWS CCP
3d printing	1	3D printing exploration

Issues:

✅ Well-structured for blog posts
❌ Not consistent with other content types (Projects, Thinking, Resources don’t use categories)
❌ “data science” is too broad (IoT vs ML vs DevOps are all lumped together)

3. Tags (Currently Used)

Blog Post Tags:

Technical: arduino, mqtt, temp sensor, python, streamlit, mermaid, github actions, jekyll
Skills: webapps, cicd, documentation, workflow
Tools: AWS, flask, dash, plotly

Other Collection Tags:

Projects/Thinking: knowledge-graph, cv, governance, nlp, rag, retrieval, analytics
Resources: visual-working-memory, data-visualization, perception, gestalt, ux-for-analytics

Issues:

✅ Rich taxonomy exists across all content
❌ Not exposed to users (no topic pages)
❌ Inconsistent formatting (knowledge-graph vs temp sensor)
❌ No cross-content discovery (can’t see all “knowledge graph” content in one place)

Proposed Phase 2 Solution: Two Complementary Improvements

IMPROVEMENT #1: Standardize Blog URL Patterns

End-State: Single, Predictable URL Pattern

Recommendation: /blog/{slug}/

Why This Pattern:

✅ Predictable: All blog posts under /blog/ prefix
✅ Clean: Simple slug-based URLs (no dates, no numbers)
✅ Conventional: Industry standard (Medium, Dev.to, Substack all use this)
✅ Flexible: Slug can include series info if needed (sensor-fleet-part-1)
✅ Breadcrumb-friendly: Home > Blog > Post Title
✅ Maintainable: One pattern forever

Migration Path

All posts become:

OLD                                          NEW
/temp-sensor-00/                        →    /blog/sensor-fleet-intro/
/temp-sensor-01/                        →    /blog/sensor-fleet-part-1/
/temp-sensor-06/                        →    /blog/sensor-fleet-part-6/
/posts/getting-started-with-github-pages/   →    /blog/getting-started-with-github-pages/
/posts/organized-mermaid-life/          →    /blog/taming-mermaid-diagrams/
/print3d-00/                            →    /blog/3d-printing-exploration/
/stock-ticker-00/                       →    /blog/stock-ticker-comparison/
/posts/earned_aws_ccp/                  →    /blog/aws-ccp-certification/

Implementation Approach:

Update _config.yml permalink pattern to /blog/:title/
Add explicit permalinks to posts that need specific slugs
Create redirect rules for old URLs (using Jekyll Redirect From plugin)
Update internal links in any posts that reference other posts
Submit old URLs to Google Search Console for remapping

Alternative Option (If You Want to Preserve Series Identity):

/blog/series/website-building/part-1/
/blog/series/sensor-fleet/intro/

Pros: Shows series structure in URL, helps with breadcrumbs
Cons: More complex, harder to type, longer URLs

My Recommendation: Stick with /blog/{slug}/

Series info can be shown in breadcrumbs and post metadata
URLs should be short and memorable
Series structure can change without breaking URLs

IMPROVEMENT #2: Create Topics Taxonomy Page

The Problem: Hidden Expertise

You have 70+ pieces of content across 5 content types (Projects, Data Stories, Thinking, Blog, Resources) but:

No way to browse by topic/expertise area
Tags exist but aren’t surfaced prominently
Users can’t discover “all knowledge graph content” or “all computer vision work”

Example User Journey (Current - Broken):

User reads “Bees, Graphs & Governance” essay
Wants to see more knowledge graph work
Has to manually click through Projects, Data Stories, Blog hoping to find related content
Gives up, bounces

Example User Journey (With Topics - Fixed):

User reads “Bees, Graphs & Governance” essay
Clicks #knowledge-graphs tag or visits /topics/
Sees: 1 Thinking piece + 2 Projects + 1 Data Story + Related blog posts
Explores deeply, stays on site

End-State: Multi-Dimensional Topic Taxonomy

Create TWO topic pages:

Option A: `/topics/` - Master Topic Index (Recommended)

A curated page showing your expertise organized by domain:

# Explore by Topic

## 🧠 Cognitive Science & Perception
- [Vision & Data Viz for Decision-Making](/thinking/vision-perception-dataviz) (Thinking)
- [Vision & Perception Cheatsheet](/resources/vision-perception-cheatsheet) (Resource)
- [7±2 is Everywhere: Chunking](/thinking/chunking) (Thinking)
- Related: `visual-working-memory`, `preattentive`, `gestalt`, `perception`

## 🕸️ Knowledge Graphs & Networks
- [Bees, Graphs & Governance](/thinking/bees-graphs-governance) (Thinking - Pinned)
- [Knowledge Network Mapping](/projects/knowledge-network-mapping) (Project - WIP)
- [Citation Link Prediction](/data-stories/citation-link-prediction) (Data Story)
- [Beehive Analytics Platform](/projects/beehive-analytics) (Project - WIP)
- Related: `knowledge-graph`, `neo4j`, `graph-neural-networks`, `metadata`

## 🤖 AI/NLP & Governance
- [RAG Without the Theater](/thinking/rag-without-theater) (Thinking)
- [Convoscope: Conversational AI](/projects/convoscope) (Project - Active)
- Related: `rag`, `retrieval`, `governance`, `nlp`, `prompting`, `prod-readiness`

## 📊 Data Products & Visualization
- [Self-Hosted Workout Intelligence](/projects/workout-intelligence) (Project - Active)
- [Exercise Dashboard: The Choco Effect](/data-stories/exercise-dashboard) (Data Story)
- [Hive Photo Metadata Tracker](/data-stories/hive-photo-tracker) (Data Story)
- Related: `analytics`, `data-visualization`, `streamlit`, `plotly`

## 🔧 DevOps & Infrastructure
- [Sensor Fleet Series](/blog/sensor-fleet-intro) (7-part blog series)
- [GitHub Actions for Jekyll](/blog/deploy-jekyll-gh-actions) (Blog)
- [Taming Mermaid Diagrams](/blog/taming-mermaid-diagrams) (Blog)
- Related: `arduino`, `mqtt`, `kafka`, `cicd`, `heroku`

## 🎓 Tutorials & Learning Guides
- [Website Building Series](/blog/getting-started-with-github-pages) (4-part series)
- [Bridge to Web Apps](/resources/bridge-to-web-apps) (Resource)
- [Project Starter Kit](/resources/project-starter-kit) (Resource)
- Related: `tutorial`, `jekyll`, `github-pages`, `streamlit`

Why This Approach:

✅ Curated: You control the narrative and groupings
✅ Cross-content discovery: Shows Projects + Thinking + Blog + Data Stories together
✅ SEO-friendly: Human-readable groupings match how people search
✅ Demonstrates breadth: Shows expertise across multiple domains
✅ Low maintenance: Manual curation means high quality
✅ Flexible: Can reorganize as your focus evolves

Option B: `/tags/` Enhancement (Complementary)

Keep the existing Jekyll tag archive but enhance it:

Add tag descriptions
Show tag relationships (parent/child)
Add “See also” suggestions
Include post count and latest activity

Example:

Tag: knowledge-graph (12 items)
Description: Graph databases, network analysis, and knowledge representation

Content:
- 2 Projects
- 3 Thinking pieces
- 1 Data Story
- 6 Blog posts

Related tags: neo4j, graph-neural-networks, metadata, cv

Implementation Approach for Topics Page

Step 1: Define Your Topic Taxonomy (Manual - One Time)

Decide on 5-7 core topics that represent your expertise:

Cognitive Science & Perception (your PhD roots)
Knowledge Graphs & Networks (systems architecture)
AI/NLP & Governance (modern AI work)
Data Products & Visualization (user-facing work)
DevOps & Infrastructure (technical implementation)
Tutorials & Learning (teaching/sharing)

Step 2: Create Topic Metadata in Front Matter (Optional - For Automation)

Add topics: field to all content:

---
title: "Bees, Graphs & Governance"
tags: [governance, knowledge-graph, metadata]
topics: [knowledge-graphs, ai-governance]  # NEW FIELD
---

Step 3: Build the Topics Page (Two Options)

Option 3A: Manual Curation (Recommended for Start)

Create _pages/topics.md
Manually list content by topic (like the example above)
Pros: Full control, beautiful curation, immediate
Cons: Requires updates when you add content

Option 3B: Automated with Liquid (For Later)

Use Jekyll’s Liquid templating to auto-generate from topics: field
Pros: No maintenance, always up to date
Cons: Requires adding topics: to all 70+ files, less control over narrative

My Recommendation: Start with Option 3A (Manual)

Faster to implement (30-60 minutes)
Better quality (you control the story)
Easier to iterate on groupings
Can automate later if needed

Cost-Benefit Analysis

Standardizing Blog URLs

Effort:

2-3 hours (update permalinks, add redirects, test links)

Value:

⭐⭐⭐⭐⭐ SEO: Better URL structure improves search rankings
⭐⭐⭐⭐ UX: Predictable URLs improve user experience
⭐⭐⭐⭐⭐ Maintenance: One pattern forever, easier to manage
⭐⭐⭐ Branding: /blog/ prefix establishes content hub

Risk:

🟡 Medium - Requires redirects, must not break existing links

ROI: HIGH - One-time effort, permanent benefit

Creating Topics Taxonomy

Effort:

Manual curation: 1-2 hours
Automated: 4-5 hours (add metadata to all files + templating)

Value:

⭐⭐⭐⭐⭐ Discoverability: Massively improves content discovery
⭐⭐⭐⭐⭐ Expertise demonstration: Shows breadth and depth of skills
⭐⭐⭐⭐ Engagement: Increases pages per session
⭐⭐⭐⭐⭐ SEO: Topic pages rank for expertise searches
⭐⭐⭐ Differentiation: Unique feature, sets you apart

Risk:

🟢 Low - Additive feature, doesn’t break anything

ROI: VERY HIGH - Moderate effort, massive ongoing benefit

Questions for You to Decide

For URL Standardization:

Pattern choice: Do you prefer…
- A) /blog/{slug}/ (simple, my recommendation)
- B) /blog/series/{series-name}/{slug}/ (preserves series structure in URL)
- C) Something else?
Slug preferences: For sensor fleet series, do you prefer…
- A) sensor-fleet-intro, sensor-fleet-part-1, etc. (descriptive)
- B) sensor-fleet-00, sensor-fleet-01, etc. (numeric, matches current)
- C) Let me auto-generate from titles
Redirect strategy:
- A) Jekyll Redirect From plugin (simple, works on GitHub Pages)
- B) Netlify/Cloudflare redirects (if you migrate hosting)
- C) HTML meta refresh as fallback

For Topics Taxonomy:

Topic groupings: Do my 6 proposed topics resonate?
- Cognitive Science & Perception
- Knowledge Graphs & Networks
- AI/NLP & Governance
- Data Products & Visualization
- DevOps & Infrastructure
- Tutorials & Learning
Would you add/remove/rename any?
Implementation approach:
- A) Manual curation first, automate later (my recommendation)
- B) Fully automated from day 1 (requires adding topics: to all files)
Topic page location:
- A) Main nav item (like Projects, Thinking, Resources)
- B) Footer only
- C) Linked from About/Resources pages
- D) All of the above
Visual treatment:
- A) Simple list (like my example above)
- B) Card grid with icons for each topic
- C) Interactive tag cloud / network visualization
- D) Minimal Mistakes archive layouts (existing theme templates)

Recommended Decision

My Opinionated Recommendation:

URL Standardization:

Pattern: /blog/{slug}/ (simple and conventional)
Slugs: Auto-generate from titles, manually tweak if needed
Series naming: sensor-fleet-intro, sensor-fleet-part-1 (human-readable)
Redirects: Jekyll Redirect From plugin

Topics Taxonomy:

Approach: Manual curation first (Option 3A)
Topics: Your 6 proposed topics are excellent, use as-is
Location: Add to footer navigation + link from About page
Visual: Start with simple list, enhance with cards later
Timeline: Implement after URL standardization (dependency: need stable URLs)

Implementation Order:

First: URL Standardization (2-3 hours)
- Provides stable foundation
- Must be done before creating cross-links in Topics page
Second: Topics Taxonomy (1-2 hours)
- Can reference new stable URLs
- Easier to curate once URLs are clean

Total Time Investment: 3-5 hours Expected Impact:

Blog traffic: +30-40% (better SEO + predictable URLs)
Pages per session: +40-50% (topic-based discovery)
Bounce rate: -25% (easier navigation between related content)

Next Steps

You review this document and answer the 7 questions above
We agree on end-state for both improvements
I implement Phase 2 based on your decisions
We test locally before pushing to production

Created: 2025-11-22 For: dagny099.github.io Phase 2 Planning Status: Awaiting user decisions before implementation

Dr. Barbara