Apply verified fact-check corrections across 12 notebooks - #11
Merged
Merged
Conversation
Applies the 85 verified corrections (and their downstream recomputations) from the eito-website fact-check audit. Highlights: - insurance: reinsurance $1.75T -> ~$0.4T; total $9.3T -> ~$8.0T (pie, icicle root + reinsurance sub-tree, prose all recomputed); stale cyber-carrier premiums (Beazley, AIG, Travelers, Chubb, Swiss Re). - cyber_insurance: US market $12.4B -> ~$9.14B (-7%); Merck v. ACE American; SolarWinds insured $0.9B -> ~$0.09B; MS Amlin $3.5B -> GBP3.5B/$5.3B; Tokio Marine est. 1879; SAP ~86% of Global Fortune 500; $9.3T -> ~$7.9T; loss-ratio series and AIG/Travelers treemap sizes corrected. - ocean: offshore-wind capacity (75->83 GW) and investment series rebuilt; Malacca 16 -> 23.7M b/d; Hormuz LNG 33% -> ~20%; shipbuilding ~90% -> ~81%; 2050 projection $8.8T -> ~$5.2T; Orsted/Hanwha revenues; adaptation-gap relabel. - actuator: Inovance revenue series de-shifted (5.6x -> 5.0x); Emerson margin. - chemical: icicle company nodes de-conflated (Henkel, Baker Hughes, Halliburton, IFF, Givaudan, BASF Care, Ecolab, Shin-Etsu, Entegris); agrochemicals label. - japan: sector shares aligned (finance/health/real estate/mfg), icicle residuals and root recomputed to JPY609.3T; treemap Toyota/TEPCO fixed, Asahi removed. - lloyds: Carbon 5757 never-launched storyline; Beazley 6500 -> ~3200 (~24x); Sukoon 2022; MCI 1966 run-off; Ki syndicate GWP. - space: launch services $15B -> $9.3B. semiconductor: EDA ~$19.2B; memory anchor. safety: Autodesk $6.1B; railway/EHS scope caveats. construction: Bouygues node. Icicle branchvalues=total consistency re-verified for every edited tree. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
Qodo reviews are paused for this user.Troubleshooting steps vary by plan Learn more → On a Teams plan? Using GitHub Enterprise Server, GitLab Self-Managed, or Bitbucket Data Center? |
The test hardcoded the pre-correction Toyota revenue (305.78) and the Mixed-remainder total (2323.38). Both were changed by the verified fact-check corrections (Toyota FY-Mar-2025 revenue $315.10B per Fortune Global 500; remainder rebalanced after adding TEPCO and removing the fabricated Asahi Mutual Life entry). Assertions now match the corrected CSV. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
![cubic-dev-ai[bot]](https://avatars.githubusercontent.com/in/1082092?s=60&v=4){kind=small}
There was a problem hiding this comment.
9 issues found and verified against the latest diff
Prompt for AI agents (unresolved issues)
Check if these issues are valid — if so, understand the root cause of each and fix them. If appropriate, use sub-agents to investigate and fix each issue separately.
<file name="notebooks/data/semiconductor_nodes.json">
<violation number="1" location="notebooks/data/semiconductor_nodes.json:60">
P2: The updated EDA parent value (19.2) does not equal its child total (13.77), so the semiconductor icicle violates the required `branchvalues="total"` invariant and can misrepresent this branch. Please reconcile the value by adding the missing child categories or setting the parent to the sum of its children.</violation>
</file>
<file name="notebooks/data/japan_top50_revenue_treemap_dataset.csv">
<violation number="1" location="notebooks/data/japan_top50_revenue_treemap_dataset.csv:2">
P2: The existing Japan economy test now fails because its CSV fixture assertions still expect the removed Toyota and remainder values. Updating the test expectations alongside this dataset change keeps CI and regression coverage aligned with the corrected data.</violation>
</file>
<file name="notebooks/insurance_industry.py">
<violation number="1" location="notebooks/insurance_industry.py:70">
P3: The modified `mo.md()` content still uses `**...**` emphasis, which can concatenate surrounding words in the rendered notebook under this project's marimo guidance. Plain text would avoid the rendering corruption across the updated note and icicle explanation.</violation>
</file>
<file name="notebooks/chemical_industry.py">
<violation number="1" location="notebooks/chemical_industry.py:127">
P1: The updated note promises a `branchvalues="total"` roll-up that the loaded chemical hierarchy cannot satisfy: `Petrochemicals & Polymers` is smaller than its direct children. Rebalancing the JSON hierarchy—such as moving company nodes under their product families or adjusting parent/child values—is needed before describing the chart as a roll-up.</violation>
</file>
<file name="notebooks/data/robotics_nodes.json">
<violation number="1" location="notebooks/data/robotics_nodes.json:54">
P3: The KION / Dematic slice can overstate comparable company revenue because €2.94B is the segment total, including €37M of intersegment revenue, rather than external revenue (€2.906B). Either use the external-revenue figure for this company slice or label the hover explicitly as total segment revenue including intersegment revenue.</violation>
</file>
<file name="notebooks/cyber_insurance_industry.py">
<violation number="1" location="notebooks/cyber_insurance_industry.py:19">
P2: The protection-gap chart presents the 2024 $9.5T figure as a 2025 value, so its 0.17% ratio is not year-consistent. Align the chart data, annotation, and title to one year, or explicitly label the ratio as a 2024/2025 comparison.</violation>
<violation number="2" location="notebooks/cyber_insurance_industry.py:324">
P2: The strategy treemap now compares AIG's US-only 2024 DWP with peers' global premium volumes, making AIG's area and the archetype totals non-comparable. Use a common geography, year, and premium basis, or separate/label the US-only figures instead of adding them to this global comparison.</violation>
<violation number="3" location="notebooks/cyber_insurance_industry.py:332">
P3: The caveat is visible only when hovering the strategy treemap; the notebook's surrounding narrative still presents the same unverified claim as established fact. Propagate the qualification to the other Trium descriptions so the exported notebook does not give conflicting confidence levels.</violation>
</file>
<file name="notebooks/chemical_nodes.json">
<violation number="1" location="notebooks/chemical_nodes.json:5">
P2: The chemicals icicle leaves roughly US$3.365T of the new US$5.5T root outside any child because its direct children total only US$2.135T. An explicit remainder node or a reconciled root value would keep the `branchvalues="total"` hierarchy consistent and prevent the chart from showing an unexplained empty portion.</violation>
</file>
Reply with feedback, questions, or to request a fix.
Re-trigger cubic
| """ | ||
| **How to read the icicle:** | ||
| - Root uses the narrower **US$3.5T** estimate to leave visual headroom; children roll up with branchvalues="total". | ||
| - Root uses the ~**US$5.5T** broad (ex-pharma) estimate; itemized segments are shown and the remainder is many smaller/formulated lines; children roll up with branchvalues="total". |
There was a problem hiding this comment.
P1: The updated note promises a branchvalues="total" roll-up that the loaded chemical hierarchy cannot satisfy: Petrochemicals & Polymers is smaller than its direct children. Rebalancing the JSON hierarchy—such as moving company nodes under their product families or adjusting parent/child values—is needed before describing the chart as a roll-up.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At notebooks/chemical_industry.py, line 127:
<comment>The updated note promises a `branchvalues="total"` roll-up that the loaded chemical hierarchy cannot satisfy: `Petrochemicals & Polymers` is smaller than its direct children. Rebalancing the JSON hierarchy—such as moving company nodes under their product families or adjusting parent/child values—is needed before describing the chart as a roll-up.</comment>
<file context>
@@ -124,7 +124,7 @@ def _(mo):
"""
**How to read the icicle:**
- - Root uses the narrower **US$3.5T** estimate to leave visual headroom; children roll up with branchvalues="total".
+ - Root uses the ~**US$5.5T** broad (ex-pharma) estimate; itemized segments are shown and the remainder is many smaller/formulated lines; children roll up with branchvalues="total".
- Petrochemicals & Polymers now shows sub-segments (ethylene, propylene, aromatics, PE, PP, PVC, PET, styrenics) with illustrative top companies under each product family.
- **Specialty Chemicals** is further broken down into key sub-segments (Paints, Adhesives, Electronics, etc.) with representative market leaders.
</file context>
| {"label": "Amkor Technology", "parent": "OSAT (assembly & test)", "value": 6.32, "hover": "2024 net sales $6.32B"}, | ||
| {"label": "JCET Group", "parent": "OSAT (assembly & test)", "value": 5.0, "hover": "2024 revenue CNY 35.96B ≈ $5.0B"}, | ||
| {"label": "EDA / ESD / IP", "parent": "Semiconductor stack (2024 est.)", "value": 17.0, "hover": "EDA/ESD market roughly $16–18B in 2024"}, | ||
| {"label": "EDA / ESD / IP", "parent": "Semiconductor stack (2024 est.)", "value": 19.2, "hover": "EDA/ESD market ~$19.2B in 2024 (ESD Alliance)"}, |
There was a problem hiding this comment.
P2: The updated EDA parent value (19.2) does not equal its child total (13.77), so the semiconductor icicle violates the required branchvalues="total" invariant and can misrepresent this branch. Please reconcile the value by adding the missing child categories or setting the parent to the sum of its children.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At notebooks/data/semiconductor_nodes.json, line 60:
<comment>The updated EDA parent value (19.2) does not equal its child total (13.77), so the semiconductor icicle violates the required `branchvalues="total"` invariant and can misrepresent this branch. Please reconcile the value by adding the missing child categories or setting the parent to the sum of its children.</comment>
<file context>
@@ -57,7 +57,7 @@
{"label": "Amkor Technology", "parent": "OSAT (assembly & test)", "value": 6.32, "hover": "2024 net sales $6.32B"},
{"label": "JCET Group", "parent": "OSAT (assembly & test)", "value": 5.0, "hover": "2024 revenue CNY 35.96B ≈ $5.0B"},
- {"label": "EDA / ESD / IP", "parent": "Semiconductor stack (2024 est.)", "value": 17.0, "hover": "EDA/ESD market roughly $16–18B in 2024"},
+ {"label": "EDA / ESD / IP", "parent": "Semiconductor stack (2024 est.)", "value": 19.2, "hover": "EDA/ESD market ~$19.2B in 2024 (ESD Alliance)"},
{"label": "Synopsys", "parent": "EDA / ESD / IP", "value": 6.13, "hover": "Fiscal 2024 revenue $6.13B"},
{"label": "Cadence Design Systems", "parent": "EDA / ESD / IP", "value": 4.64, "hover": "2024 revenue $4.64B"},
</file context>
| @@ -1,5 +1,5 @@ | |||
| Rank,Company,Type,RevenueUsdBillions | |||
| 1,Toyota,Public,305.78 | |||
| 1,Toyota,Public,315.10 | |||
There was a problem hiding this comment.
P2: The existing Japan economy test now fails because its CSV fixture assertions still expect the removed Toyota and remainder values. Updating the test expectations alongside this dataset change keeps CI and regression coverage aligned with the corrected data.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At notebooks/data/japan_top50_revenue_treemap_dataset.csv, line 2:
<comment>The existing Japan economy test now fails because its CSV fixture assertions still expect the removed Toyota and remainder values. Updating the test expectations alongside this dataset change keeps CI and regression coverage aligned with the corrected data.</comment>
<file context>
@@ -1,5 +1,5 @@
Rank,Company,Type,RevenueUsdBillions
-1,Toyota,Public,305.78
+1,Toyota,Public,315.10
2,Honda,Public,142.45
3,Mitsubishi Corporation,Public,133.59
</file context>
| {"company": "Trium Cyber", "country": "UK", "strategy": "Technical underwriters", "premium_b": 0.10, | ||
| "focus": "specialist", "cyber_pct": 100, | ||
| "hover": "First Lloyd's-approved mono-line cyber syndicate (1322). 100% cyber. £73M premium, 375 policies, ZERO claims in 3 years."}, | ||
| "hover": "First Lloyd's-approved mono-line cyber syndicate (1322). 100% cyber. £73M premium, 375 policies, ZERO claims in 3 years (company-stated; not independently verified)."}, |
There was a problem hiding this comment.
P3: The caveat is visible only when hovering the strategy treemap; the notebook's surrounding narrative still presents the same unverified claim as established fact. Propagate the qualification to the other Trium descriptions so the exported notebook does not give conflicting confidence levels.
Prompt for AI agents
Check if this issue is valid — if so, understand the root cause and fix it. At notebooks/cyber_insurance_industry.py, line 332:
<comment>The caveat is visible only when hovering the strategy treemap; the notebook's surrounding narrative still presents the same unverified claim as established fact. Propagate the qualification to the other Trium descriptions so the exported notebook does not give conflicting confidence levels.</comment>
<file context>
@@ -321,15 +321,15 @@ def _(go):
{"company": "Trium Cyber", "country": "UK", "strategy": "Technical underwriters", "premium_b": 0.10,
"focus": "specialist", "cyber_pct": 100,
- "hover": "First Lloyd's-approved mono-line cyber syndicate (1322). 100% cyber. £73M premium, 375 policies, ZERO claims in 3 years."},
+ "hover": "First Lloyd's-approved mono-line cyber syndicate (1322). 100% cyber. £73M premium, 375 policies, ZERO claims in 3 years (company-stated; not independently verified)."},
{"company": "Hiscox", "country": "UK", "strategy": "Technical underwriters", "premium_b": 0.38,
"focus": "diversified", "cyber_pct": 7.5,
</file context>
- chemical_nodes.json: raise 'Petrochemicals & Polymers' parent to $875B so it covers its overlapping product-market children (branchvalues=total requires parent >= children; was 700 < 872); hover notes it's a gross sum. - ocean: wire Strait of Malacca throughput into the chokepoint hover (was collected but never displayed); attribute the 5th-largest ranking to OECD only; caveat that the 2050 trajectory bars blend different metrics. - lloyds: recolour MCI 1966 timeline marker to run-off (was still active/green). - cyber_insurance: align the CrowdStrike appendix insured range (~$0.54-1.08B) and the Japan-market tooltip ranking with the corrected prose. - Escape $ in the mo.md() cells edited by the fact-check corrections (insurance, semiconductor, chemical, actuator, safety, space, robotics), per the notebooks' marimo rendering convention. False positives left as-is: EDA ($19.2B) and chemical root ($5.5T) headroom over child nodes is the notebooks' disclosed 'company nodes don't sum to parents' design (children < parent renders fine). Japan CSV test already updated in e5d939b. Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
2 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
Applies 85 verified fact-check corrections (and their downstream recomputations) across 12 marimo notebooks. Each discrepancy was independently re-verified against a credible primary/authoritative source before editing — 82 confirmed, 3 partial (verified value used), 8 refuted and left unchanged.
Corrections by notebook
$0.4T (pie + icicle root/reinsurance sub-tree + all "$9.3T" prose → ~$8.0T recomputed); stale cyber-carrier premiums (Beazley $1.276B, AIG ~$0.28B, Travelers ~$0.54B, Chubb/Swiss Re re-labeled as estimates); source re-attributed to Allianz.90%→~81%; 2050 $8.8T→**$5.2T**; Equinor 10–12 GW; Ørsted $10.3B; Hanwha Ocean $7.9B; adaptation-gap relabel (all-sector); dry bulk basis; shipping-vs-aviation; Suez rerouting; "7th"→"5th largest"; UNCTAD $2.2T = ocean trade.12%→19%; Nabtesco FY self-contradiction; HVAC scope caveat.$17B → **$19.2B** + memory anchor; stack $805B→$807B. safety Autodesk $6.1B (FY2025); railway/EHS scope caveats. construction Bouygues node €10.34B ≈ $11.2B.Verification
branchvalues="total"trees re-verified (parent = sum of children); Japan root and insurance reinsurance sub-tree recomputed exactly./notebookspage).🤖 Generated with Claude Code
Generated by Claude Code
Summary by cubic
Applies 85 verified fact corrections across 12 notebooks, recomputes affected charts so values and prose align with primary sources, and updates tests to match the corrected Japan treemap.
$0.54–1.08B; MS Amlin £3.5B ($5.3B); SAP ~86% GF500; clarified Lloyd’s silent-cyber timeline and Japan life ranking; adjusted loss-ratio series.Written for commit 05ed290. Summary will update on new commits.