Presenting Findings - Delivery Constraint & Opportunity Analysis

Description

Turn proven impact findings into one stakeholder story: reconstructed cycle first, then the matrix, then stage detail. Render with Jupyter, Streamlit, QuickSight, or Power BI (skill pick). Keep current through Huginn until a connector exists.

Activities in this Phase

Order	Name	Actions
Tell the Data Story
#1	Design the Big-Picture-to-Detail Narrative	Read the 'TSY-1 Design the Big-Picture-to-Detail Narrative' activity from playbook 'Delivery Constraint & Opportunity Analysis' (workflow 'Tell the Data Story'). Follow its guidance when assisting me. ## Purpose Plan the stakeholder story: reconstructed cycle first (three metrics), then quadrant, then the deficiency register, then stage detail, then starter tabs. Never lead with a detail chart. Do not re-classify Action Register tags. ## Inputs - YAML in `streamlit/data/` (cost_of_delay, ai_opportunities, action_register, promoted_eda). - Template `docs/analysis/narrative-outline.md`. - Skill Data Storytelling and Dashboard Design → Recipe. - Skill Metric Render Catalog (optional pairings; Streamlit tab map). ## Steps 1. Open `docs/analysis/narrative-outline.md`. 2. Opening visual: Delivery cycle (effort, wall-clock, impact). Then opportunity quadrant. Then deficiency register (finding + decision + dE/dK/dI + receiver). Copy CAS; do not invent tags. - dE + automate/delegate → AI SDLC (receiver only). - Other dE → process/tooling. - dK → Team Upskill Engine input (stage, clock, unit, n, role). - dI → staffing. 3. One section per stage in flow order: finding sentence + Action Register decision (or no action). Embed promoted-eda ids or "none". 4. Cap one visual per finding. Name the clock and n. Do not equate issue business days with MR hours. 5. People: name the role (author / integrator / reviewer). Concentration is a risk. 6. If Present Findings uses Streamlit or Jupyter, list starter tabs to instantiate or skip: Commits, Work items, Changes, Pipelines, Load, Defects, Sprints (skill Present Findings with Streamlit → Starter mart tabs). Do not put a tab first. 7. Name the presentation skill (Jupyter = `notebooks/published-findings.ipynb`; Streamlit = mart pages 0–10). ## Done when - Outline lists cycle, quadrant, register, stage sections, and tabs-to-skip. - Duration KPIs state unit, cohort, and n. - Register beat names receivers only. ## Feeds Present Findings.
#2	Present Findings	Read the 'TSY-2 Present Findings' activity from playbook 'Delivery Constraint & Opportunity Analysis' (workflow 'Tell the Data Story'). Follow its guidance when assisting me. ## Purpose Build the stakeholder story from the outline. Pick one presentation skill. Do not re-derive numbers. Do not re-classify tags. ## Inputs - `docs/analysis/narrative-outline.md`. - `streamlit/data/.yaml`. - Skills: Present Findings with Jupyter* (`notebooks/published-findings.ipynb`), Present Findings with Streamlit (pages 0–10), QuickSight, Power BI. - Skill Metric Render Catalog. ## Steps 1. Choose one skill. Skip QuickSight if unsubscribed. Skip Power BI if no tenant. Do not add BI in CDK. 2. Order: Delivery cycle → Opportunity quadrant → Deficiency register → Stage detail (promoted_eda only) → starter tabs. 3. Streamlit: pages already in the starter (`0_Delivery_Cycle.py` … `10_Sprints.py`). Browser-verify the path end to end, then each instantiated tab. Skip tabs whose contract views are missing (`tab_pages.py`). If the view exists, run Athena via shared `streamlit/tab_athena.py` — do not leave `print("Instantiate: …")` as the tab body. Pin QueryExecutionId; empty cohort names the window. Live GitHub-only lake: instantiate Commits, Work items (volume only — omit lead/cycle without `jira_status_segments`), Changes (MR hours, not issue days), Pipelines (top failing = workflow name; job-level extract is not in the lake), Load (risk signal). Skip Defects and Sprints unless those views exist. 4. Jupyter: open `notebooks/published-findings.ipynb`, kernel Python (opsdl .venv) (`python -m ipykernel install --user --name opsdl --display-name "Python (opsdl .venv)"`; notebook root = repo). Restart & Run All. Skill 114 recipe matches the cell order. 5. Receivers named only: AI SDLC, process/tooling, Team Upskill Engine (candidates), staffing. No TUE modules. No agent names. 6. Write Published Findings Notebook/Dashboard = the mart URL and/or the notebook path. ## Pitfalls - If `LoginRefreshRequired` / `CreateOAuth2Token` expired → `aws login`, restart kernel, not a Glue/Athena ACL bug. - `streamlit/data/promoted_eda.yaml` `rows: []` is still correct until activity 398 promotes Look-until-Trusted survivors. TSY floor is still cycle + quadrant + deficiency register. Do not dump Metric Render Catalog. Do not treat empty promoted-eda as a bug in this activity. ## Done when - A stakeholder can follow cycle → quadrant → register without a debugging session. - Numbers match the YAML ranking files. - Tags match CAS. Omitted tabs listed as skipped for missing views. ## Feeds Configure Huginn for Automated Reporting. ## Skip Do not open with a single-stage chart or Commits tab. Do not name agents to implement.
#3	Configure Huginn for Automated Reporting	Read the 'TSY-3 Configure Huginn for Automated Reporting' activity from playbook 'Delivery Constraint & Opportunity Analysis' (workflow 'Tell the Data Story'). Follow its guidance when assisting me. ## Purpose Keep the published story's underlying data current. Huginn has no MCP connector today. ## Inputs - Published Findings (mart URL or notebook). - Skill Huginn Browser Automation → Recipe. - Starter `docs/huginn/README.md`. ## Steps 1. Read `docs/huginn/README.md`. Check sibling `../huginn` for an existing Agent/Scenario. 2. Huginn UI URL is org-specific — record it in playbook notes. Search Scenarios for lake / Athena / Streamlit. 3. Agent type follows the Present Findings surface (HTTP GET of `streamlit_mart_url` from `docs/deploy-manifest.yaml` if Streamlit). Jupyter has no live URL — defer if that was the surface. 4. Dry-run once. Record Agent name, schedule, purpose. 5. Flag as a future MCP-connector candidate. ## Done when - A scheduled refresh exists, or an explicit deferral is written in `docs/huginn/README.md`. ## Skip when `streamlit_mart_url` is empty — defer and say so.