DSRS Knowledge-Base Prototype

What’s inside
Prototype KB (dsrs_kb_source.csv), a NetworkX graph converted to an interactive PyVis HTML (dsrs_graph.html), and a fully-reproducible notebook (DSRS_KB_Prototype.ipynb).

How to run

1. Open DSRS_KB_Prototype.ipynb and run Runtime ▸ Run all
2. When execution finishes, double-click dsrs_graph.html in the file browser to explore the graph.

Key files

File	Purpose
dsrs_kb_source.csv	Source rows for the knowledge-base (23 rows)
dsrs_graph.html	Interactive graph (23 nodes / 20 edges)
DSRS_KB_Prototype.ipynb	Notebook that builds the CSV, graph and demo queries

Sources
All KB rows cite:

• DSRS Strategy.pdf
• Public pages on dsrs.illinois.edu
• LinkedIn post permalinks (last 6 months)

Maintenance & Growth Strategy

1 . Keeping the KB current

Task	Owner	Frequency	Tooling
Automated scrape of DSRS website and LinkedIn (new posts, new datasets) → append to dsrs_kb_source.csv	Infrastructure	nightly cron on DSRS server	Python + GitHub Actions
CSV diff review: check duplicates / stale rows flagged by last_updated > 90 days	Sub-unit leads	monthly	notebook validator
Manual additions (new projects, workshop decks)	Services & Data Hub staff	ad-hoc	KB template form (Google Form → Zapier → CSV row)

2 . Accuracy controls

• All rows include a source URL/PDF; rows without a source fail the CI check.
• last_updated timestamp is overwritten on every save so ageing rows are obvious.
• Graph build step refuses to run if duplicate id values are detected.

3 . Integration with day-to-day workflows

• New project kickoff → PM completes a 5-line KB form; CI merges it automatically.
• LinkedIn post → Zapier webhook auto-creates a news row with post_date.
• Dataset on-boarding in Data Hub playbook now ends with “add dataset to KB” step.
• Monthly DSRS all-hands starts with the KB growth chart so gaps are visible.

4 . Road-map (next 6 months)

1. Migrate CSV to a lightweight Postgres table exposed via REST.
2. Build a Neo4j mirror for advanced graph queries.
3. Embed a Streamlit KB search app on the DSRS intranet.
4. Add unit tests to guarantee schema consistency before every merge.