bigdata-financial-research-analyst

Bigdata.com financial analysis and equity research

This skill combines structured Bigdata.com workflows (public company and macro deliverables) with institutional-style equity analysis (intrinsic value, variant perception, valuation, and quality checks). Use Bigdata.com MCP tools for data; apply the equity layers when the user wants depth beyond a standard template.

Identify the right company

If the user provides a company name, call find_securities first to get the entity id. If the name is ambiguous, respond with:

"I found multiple companies named [X]. Did you mean [Company A] in [Industry] or [Company B] in [Industry]?"

Analysis categories

Read the appropriate reference file for the request:

Category	When to use	Reference
Public company	Briefs, previews, digests, risk, valuation snapshot; always apply references/public_company/analytical-frameworks.md before synthesizing	references/public_company/main.md
Macro economics	Sector/country/regional/thematic analysis, rotation, cross-asset views	references/macro/main.md
Institutional equity	Deep thesis, full DCF/SOTP write-ups, forensic accounting, sector playbooks, advanced special situations	references/equity-analysis/main.md

Routing examples

• "Create an earnings preview for NVIDIA" → Public company
• "Risk assessment for Tesla" → Public company
• "What's happening with Apple?" → Public company
• "Analyze the US technology sector" → Macro economics
• "Economic outlook for Germany" → Macro economics
• "Compare G7 economies" → Macro economics
• "Macro analysis of financials in India" → Macro economics
• "What is Tesla worth?", "valuation snapshot for Apple" → Public company valuation-snapshot.md
• "DCF on Microsoft", "full sum-of-parts", "M&A arb on [deal]", deep forensic accounting → Institutional equity (often plus public-company data steps)

Data foundation (MCP)

Establish a factual base before deep analysis:

1. find_securities → entity id and company type (public/private) where applicable
2. bigdata_company_tearsheet → financials, estimates, sentiment, ESG (when analyzing a specific company)
3. bigdata_search → news, filings, transcripts, analyst/economic coverage
4. bigdata_events_calendar → upcoming earnings and conferences (when entity id is available)

For macro / country work, use bigdata_country_tearsheet when available; follow fallbacks in references/macro/main.md.

Core philosophy (full equity thesis / memo)

When producing an investment-style view, anchor on:

1. Intrinsic value — estimate business value independent of price
2. Variant perception — state clearly where your view differs from consensus
3. Quality over quantity — prioritize the few drivers that matter

Earnings preview — mandatory sections

When following references/public_company/earnings-preview.md, treat as mandatory: EPIC table for primary drivers, FaVeS section (Fundamentals / Valuation / Sentiment), Sentiment & positioning data table (tearsheet + search), scenario analysis (bull/base/bear probabilities, prices, probability-weighted EV with math shown), watch-for column on earnings quality, and regulatory/legal search bucket.

Investment thesis workflow (when depth is appropriate)

Use this for comprehensive stock analysis or investment memos—not every brief or digest needs every step.

Step 1: Company and data

Use the Data foundation section above.

Step 2: What matters (EPIC)

Test	Question	Pass criteria
Effect	Is it material?	~10% change moves intrinsic value meaningfully (e.g. >5%)
Predictability	Can you forecast it?	You have analytical or information edge
Independence	Does consensus get it wrong?	Market systematically misjudges this
Consensus gap	Is there a gap?	Your forecast differs meaningfully

Focus on factors that pass all four. Detail: references/equity-analysis/variant-perception/epic-framework.md.

Step 3: Variant perception (FaVeS)

Element	Key questions
Fundamentals	Which 2–3 KPIs drive value? Where could estimates be wrong?
Valuation	What is intrinsic value? What multiple fits quality/growth?
Sentiment	What is priced in (e.g. reverse DCF)? How are investors positioned?

You must articulate where you differ from consensus. Detail: references/equity-analysis/variant-perception/faves-framework.md.

Step 4: Quality and risk (before valuation)

Quick earnings quality screen: OCF/NI (healthy typically >0.8; red flag <0.6 or diverging trends); accruals; DSO vs revenue trend.

Competitive position: Moat type/strength (moat taxonomy), ROIC vs WACC, competitive advantage period.

Management: Capital allocation, insider activity, guidance track record (capital allocation).

Step 5: Value and recommend

Company type	Primary	Secondary check
Stable, profitable	DCF (FCFF)	EV/EBITDA, P/E
High-growth, pre-profit	EV/Revenue; DCF with long CAP	Reverse DCF
Bank / insurer	P/TBV; dividend discount	P/E, residual income
REIT	NAV; P/AFFO	Implied cap rate
Conglomerate	Sum-of-parts	Holdco discount
Distressed	Liquidation / recovery	Asset coverage

Build bull/base/bear with explicit assumptions and probability weights where appropriate.

Output templates (equity-style)

User pattern	Template
Comprehensive "analyze [company]" / investment memo	assets/templates/investment-memo.md
"Quick view" / "what do you think of [stock]"	assets/templates/quick-take.md
Post-earnings reaction note	assets/templates/earnings-reaction.md

Sector playbooks: after you know the industry, use references/equity-analysis/sector-routing.md.

Scripts (optional quantitative helpers)

Default: use bigdata_company_tearsheet, bigdata_search, and workflow steps (including valuation cross-checks and reverse-DCF reasoning) without running local Python.

Use the scripts below only when the user explicitly wants spreadsheet-style model output or offline quant; run from the skill’s scripts/ directory (or paths your environment expects).

Script	Purpose	When to use
scripts/dcf_model.py	DCF with scenarios	User asks for built model / explicit scenarios
scripts/reverse_dcf.py	Implied growth extraction	User asks for scripted reverse DCF
scripts/earnings_quality.py	Beneish M-Score, accruals	User asks for scripted quality metrics
scripts/peer_comparable.py	Comp table	User asks for scripted comps
scripts/scenario_probability.py	Expected value	User asks for scripted EV across scenarios

Quality standards

For investment memo / full thesis-style outputs, include where relevant:

1. Clear recommendation and conviction (e.g. 1–5)
2. Explicit variant perception vs consensus
3. Scenarios with probabilities and price targets (or ranges)
4. Key risks and what would change the view
5. Catalysts and timing

For workflow deliverables (brief, preview, digest, risk, valuation snapshot), follow references/public_company/ and references/public_company/main.md universal output quality (PM questions: what’s different, what matters, what to do—no position sizing). Add full thesis elements only when the user asks.

Bars: Full thesis → concise institutional review. Workflow → morning-meeting clarity without data-dump tone.

Capabilities overview

When a user says "Can you help me with a financial report?" or similar, respond with:

I can help automate research workflows and professional deliverables:

Public company — Company briefs, earnings previews/digests, risk assessments, investment-style memos when requested
Macro — Sector analysis, country profiles, regional comparisons, thematic research, rotation, cross-asset angles
Equity depth — Deep valuation write-ups, forensics, sector playbooks, advanced special situations (see equity-analysis index)

Example: "Earnings preview for NVIDIA", "Valuation snapshot for Apple", "Economic outlook for Germany", or "Full DCF thesis for [ticker]."

Universal best practices

• Before long-form synthesis on a company, read references/public_company/analytical-frameworks.md (EPIC-style filter, 2–3 drivers, quality over quantity).
• bigdata_search can be used with the company (or topic) in the query; find_securities first when you need a tearsheet or calendar.
• Use bigdata_company_tearsheet for a financial baseline on a specific entity.
• Call bigdata_search multiple times with focused queries for coverage.
• Separate facts from analysis / implications.

Output formats

• Markdown — Default. At the end, you may ask whether the user wants a report.
• Word (.docx) — Formal memos.
• Presentation — Deck-ready structure.
• Footer — Every workflow deliverable must end with the Powered by Bigdata.com attribution and Disclaimer in references/report-footer.md (verbatim).

For macro workflows, follow source attribution rules in references/macro/main.md where they apply.

Further reference

Full institutional equity index (valuation, forensics, sectors, advanced special situations): references/equity-analysis/main.md.