Skip to content

level-5: Ankit Kumar Singh#535

Open
ankitsinghh007 wants to merge 2 commits into
Life-Atlas:masterfrom
ankitsinghh007:level-5-submission
Open

level-5: Ankit Kumar Singh#535
ankitsinghh007 wants to merge 2 commits into
Life-Atlas:masterfrom
ankitsinghh007:level-5-submission

Conversation

@ankitsinghh007
Copy link
Copy Markdown
Contributor

Submission Level

Level: 5


What I Did

Completed the Level 5 Graph Thinking challenge using the provided factory datasets.

Included Work

  • Knowledge graph schema design
  • Node and relationship modeling
  • Cypher query reasoning
  • SQL vs graph traversal comparison
  • Bottleneck and capacity analysis
  • Hybrid vector + graph reasoning concepts
  • Level 6 architecture planning
  • Detailed HOW_I_DID_IT reflection

Focus Areas

  • Factory operational relationships
  • Connected system modeling
  • Knowledge graph thinking
  • Industrial digital twin concepts
  • Neo4j and Cypher reasoning

Checklist

  • I read the README and CONTRIBUTING guide
  • My PR title follows the format: level-X: Your Name
  • I tested my changes locally before submitting

Signed-off-by: Ankit Kumar Singh

Copilot AI review requested due to automatic review settings May 13, 2026 18:06
Copy link
Copy Markdown

Copilot AI left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Pull request overview

This PR adds a Level 5 written submission for graph-thinking around the factory production dataset, including schema design, query reasoning, bottleneck analysis, hybrid vector/graph concepts, and a Level 6 dashboard plan.

Changes:

  • Adds Level 5 answers covering all five challenge questions.
  • Adds a Mermaid-based factory knowledge graph schema.
  • Adds a reflection document describing the author’s process.

Reviewed changes

Copilot reviewed 3 out of 3 changed files in this pull request and generated 7 comments.

File Description
submissions/ankitsinghh007/level5/answers.md Main Level 5 written response with schema, SQL/Cypher examples, bottleneck analysis, hybrid query, and L6 plan.
submissions/ankitsinghh007/level5/schema.md Mermaid schema diagram and node/relationship count summary.
submissions/ankitsinghh007/level5/HOW_I_DID_IT.md Reflection on the approach and lessons learned.

💡 Add Copilot custom instructions for smarter, more guided reviews. Learn how to get started.

| HAS_CAPACITY | 8 | one per week |
| TRIGGERED_AT | ~4 | bottleneck nodes |
| **TOTAL** | **~193** | |
```
- w6: deficit **−80 hrs** — capacity 440, planned 520
- w7: deficit **−80 hrs** — capacity 520, planned 600

**Projects/stations causing w1 overload (planned hours > 40, actual > planned):**
Comment on lines +206 to +208
- Station 016 (Gjutning): P03 w2 +25%, P05 w2 +14.3%, P07 w2 +10%, P08 w3 +13.6%
- Station 014 (Svets o montage): P03 w1 +14.3%, P04 w1 +12%, P05 w1 +6.9%, P08 w1 +10%
- Station 021 (SR B/F-hall): P04 w2 +8.3%
b.severity = CASE WHEN excess > 20 THEN "HIGH"
WHEN excess > 10 THEN "MEDIUM"
ELSE "LOW" END
MERGE (b)-[:TRIGGERED_AT]->(s)

Filtering by product type (`IQB`, `IQP`, `SB`) finds projects that made the same things — but a hospital project and a parking structure both use IQB beams. The hybrid approach finds projects that were *operationally similar*: same stations (same production flow), similar complexity, AND historically low variance. That last condition is the key — it surfaces past projects whose actual execution matched their plan, which are the only useful analogues for estimating future schedules. A product-type filter has no way to surface execution quality from historical data.

**Boardy parallel:** Instead of matching by project scope, Boardy matches people whose embedded `needs` description is close in vector space to another person's `offers` description, AND who are in the same community in the relationship graph (KNOWS, ATTENDS, WORKS_AT edges). The vector surface plausible matches; the graph confirms they're socially reachable. Neither alone would work as well.
| `(Week)-[:HAS_CAPACITY]->(Capacity)` | each capacity row | `deficit`, `total_capacity`, `total_planned` |
| `(Station)-[:HAD_BOTTLENECK]->(Week)` | computed: actual > planned * 1.1 | `severity`, `excess_hours` |

### 4 Streamlit Dashboard Panels
Copy link
Copy Markdown
Contributor Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

The “4 Streamlit Dashboard Panels” heading refers to Section 4 of the document, not the total panel count.

Comment on lines +259 to +270
MATCH (new_project:Project {id: "P05"})
MATCH (new_project)-[:SCHEDULED_AT]->(shared_station:Station)
MATCH (candidate:Project)-[r:SCHEDULED_AT]->(shared_station)
WHERE candidate.id IN similar_project_ids
AND candidate.id <> new_project.id
WITH candidate,
count(DISTINCT shared_station) AS shared_stations,
avg(abs(r.actual_hours - r.planned_hours) / r.planned_hours) AS avg_variance
WHERE avg_variance < 0.05
RETURN
candidate.name AS similar_project,
shared_stations,
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

2 participants