Hugging Face's logo

Datasets:
speb
/
financial-data

Modalities:
Text
Formats:
parquet
Size:
100M - 1B
Libraries:
Datasets
Dask
Polars
License:
Dataset card Data Studio Files
xet
 Community
1
financial-data / README.md
speb's picture
speb
Upload folder using huggingface_hub
53ceeae verified
preview code
|
Raw
History Blame Contribute Delete
40.8 kB
metadata 
license: cc-by-4.0

Financial Data — Marts Schema Data Dictionary

This document provides a comprehensive schema reference and metric dictionary for the 23 analytical tables compiled in the marts schema of database.db (and saved as Parquet files under marts/).

Table of Contents

  1. fct_combined_scorecard (Unified screening table)
  2. dim_company_profiles (Company metadata)
  3. fct_financial_ratios (Core ratios, owner earnings, RONTA)
  4. fct_growth_rates (Multi-year CAGRs & consistency)
  5. fct_earnings_quality (Sloan accruals & Beneish M-Score)
  6. fct_dupont_decompositions (3-factor and 5-factor ROE decompositions)
  7. fct_valuation_multiples (Daily historical multiples & Z-Scores)
  8. fct_macro_sensitivity (Treasury beta & R-squared)
  9. fct_insider_sentiment (Rolling 30d/90d/180d net transactions)
  10. fct_dividend_history (Chronological distributions & streaks)
  11. fct_magic_formulas (Greenblatt ranking)
  12. fct_piotroski_fscores (9-point financial health scorecard)
  13. fct_altman_zscores (Credit strength & solvency)
  14. fct_shareholder_yields (Total capital return yield)
  15. fct_momentum_metrics (Relative price strength)
  16. fct_lynch_categories (Peter Lynch stock archetypes)
  17. fct_screener_recommendations (Long/Short composite scorecard)
  18. fct_capital_allocation (ROIC & incremental capital efficiency)
  19. fct_dcf_valuations (WACC & multi-scenario DCF price targets)
  20. fct_sector_benchmarks (Sector and industry percentile benchmarks)
  21. fct_macro_rates (Daily Treasury yields & exchange rates)
  22. fct_earning_call_transcripts (Quarterly call text paragraphs)
  23. fct_sec_filing_sections (SEC MD&A & Risk Factors text sections)

1. fct_combined_scorecard

  • Purpose: The "front page" dashboard. Flat, denormalized view containing the latest values of every calculated metric joined into a single table for fast querying.
  • SQL Source: Derived from private transformation 9_z_combined_scorecard.sql (which ingests: dim_company_profiles, fct_financial_ratios, fct_growth_rates, fct_earnings_quality, fct_dupont_decompositions, fct_valuation_multiples, fct_macro_sensitivity, fct_insider_sentiment, fct_dividend_history, fct_magic_formulas, fct_piotroski_fscores, fct_altman_zscores, fct_shareholder_yields, fct_momentum_metrics, fct_lynch_categories, fct_screener_recommendations, fct_capital_allocation, fct_dcf_valuations, and fct_sector_benchmarks from the marts schema; statements_usd from the staging schema; and stock_valuation_snapshot from the raw schema).
  • Key Columns:
    • symbol (VARCHAR): Stock ticker symbol (e.g. AAPL).
    • report_date (VARCHAR): Date of the latest statement period.
    • market_cap (DOUBLE): Market capitalization in USD.
    • pe_ratio (DOUBLE): Price-to-Earnings ratio.
    • altman_z_score (DOUBLE): Altman Z-score.
    • piotroski_f_score (INTEGER): Piotroski F-score (0 to 9).
    • magic_formula_rank (BIGINT): Overall rank according to the Magic Formula.
    • total_shareholder_yield_pct (DOUBLE): Combined yield (dividends, buybacks, debt paydowns).
    • composite_momentum_pct (DOUBLE): Weighted price momentum score.
    • long_short_recommendation (VARCHAR): Recommended screener classification (STRONG_BUY, BUY, NEUTRAL, SHORT, STRONG_SHORT).
    • rev_cagr_3y (DOUBLE): 3-year Revenue CAGR.
    • accrual_ratio (DOUBLE): Sloan accruals ratio.
    • beneish_m_score (DOUBLE): Beneish M-Score.
    • roe_archetype (VARCHAR): DuPont ROE driver archetype.
    • pe_z_score (DOUBLE): Standard deviations of PE vs. historical mean.
    • treasury_beta_1y (DOUBLE): Rolled 1-year regression beta vs. 10Y US Treasury yield change.
    • insider_rolling_90d_net_value (DECIMAL): USD net buy/sell transactions by corporate insiders.
    • dividend_streak_years (BIGINT): Consecutive years of dividend increases.

2. dim_company_profiles

  • Purpose: Holds static and semi-static qualitative metadata for each corporation.
  • SQL Source: Derived from private transformation 1.8_dim_company_profiles.sql (which ingests: stock_profile from the raw schema).
  • Key Columns:
    • symbol (VARCHAR): Primary Key.
    • sector (VARCHAR): Macro sector (e.g. Technology).
    • industry (VARCHAR): Micro industry (e.g. Consumer Electronics).
    • long_business_summary (VARCHAR): Full textual description of company operations.
    • full_time_employees (BIGINT): Current headcount.
    • web_site (VARCHAR): Corporate URL.

3. fct_financial_ratios

  • Purpose: Core financial analysis metrics, normalized balance sheet totals, and specialized custom metrics (RONTA, capex-to-operating-cash-flow, R&D-adjusted asset values).
  • SQL Source: Derived from private transformation 2_financial_ratios.sql (which ingests: dim_company_profiles from the marts schema; prices and statements_usd from the staging schema; and stock_analyst_price_targets, stock_shares_outstanding, and stock_valuation_snapshot from the raw schema).
  • Key Columns:
    • net_tangible_assets (DOUBLE): Total Assets - Total Liabilities - Goodwill - Intangible Assets.
    • owner_earnings (DOUBLE): Net Income + D&A - Capital Expenditures.
    • ronta_pct (DOUBLE): Return on Net Tangible Assets. Buffett's favorite profitability metric.
    • buffett_classification (VARCHAR): Classification of moat quality (The Great, The Good, The Gruesome).
    • analyst_consensus_upside_pct (DOUBLE): Percentage difference between current stock price and mean analyst price target.
    • interest_coverage_ratio (DOUBLE): EBIT / Interest Expense.
    • rule_of_40_score (DOUBLE): Revenue Growth % + Free Cash Flow Margin % (used for software/SaaS business models).
    • is_excluded (BOOLEAN): Flag denoting if the company belongs to financial/cyclical industries that distort structural ratio calculations.

4. fct_growth_rates

  • Purpose: Multi-year compound annual growth rates (CAGRs) for major items on the income and cash flow statements, along with consistency and momentum checks.
  • SQL Source: Derived from private transformation 9.0_growth_rates.sql (which ingests: statements_usd from the staging schema).
  • Key Columns:
    • rev_growth_1y / net_growth_1y / fcf_growth_1y (DOUBLE): YoY growth rates.
    • rev_cagr_3y / net_cagr_3y / fcf_cagr_3y (DOUBLE): 3-year compound annual growth rates.
    • rev_cagr_5y / net_cagr_5y / fcf_cagr_5y (DOUBLE): 5-year compound annual growth rates.
    • rev_deceleration_flag (BOOLEAN): True if current YoY growth is lower than the previous period's growth.
    • rev_growth_consistency_score (DOUBLE): Standard deviation of growth rates over the lookback window (lower = more predictable growth).

5. fct_earnings_quality

  • Purpose: Evaluates whether reported net income is backed by cash flow, and flags signs of earnings manipulation or accounting anomalies.
  • SQL Source: Derived from private transformation 9.1_earnings_quality.sql (which ingests: statements_usd from the staging schema).
  • Key Columns:
    • accrual_ratio (DOUBLE): Sloan accruals metric: (Net Income - FCF) / Total Assets.
      • Threshold: Values > 0.10 indicate excessive accruals (earnings ahead of cash).
    • fcf_conversion_rate (DOUBLE): FCF / Net Income.
      • Interpretation: Ideal is > 1.0. Sustained values < 0.70 suggest weak earnings quality.
    • beneish_m_score (DOUBLE): Mathematical model utilizing 8 accounting indexes (DSRI, GMI, AQI, SGI, DEPI, SGAI, LVGI, TATA) to detect earnings manipulation.
      • Threshold: Scores > -1.78 denote possible manipulation; > -1.49 represents a high-risk manipulation signal.
    • beneish_m_score_label (VARCHAR): Rating of manipulation risk (High Risk, Safe).

6. fct_dupont_decompositions

  • Purpose: Breaks down Return on Equity (ROE) into 3-factor and 5-factor component parts to determine if profitability is driven by profit margins, asset efficiency, or financial leverage.
  • SQL Source: Derived from private transformation 9.15_dupont_decomposition.sql (which ingests: statements_usd from the staging schema).
  • Key Columns:
    • roe_pct (DOUBLE): Profit / Common Equity.
    • net_margin_pct (DOUBLE): Profit margin (Net Income / Revenue).
    • asset_turnover (DOUBLE): Asset efficiency (Revenue / Average Assets).
    • equity_multiplier_leverage (DOUBLE): Financial leverage (Average Assets / Average Equity).
    • ebit_margin_pct (DOUBLE): Operating profitability.
    • interest_burden_factor (DOUBLE): Pretax Income / EBIT.
    • tax_burden_factor (DOUBLE): Net Income / Pretax Income.
    • roe_archetype (VARCHAR): Business return driver label:
      • High-Margin Compounder (High profit margin, low leverage)
      • Asset-Light Asset Turner (High asset turnover, low leverage)
      • Leverage-Driven Returns (Low margins/turnover, high leverage)
      • Underperforming (Negative ROE)

7. fct_valuation_multiples

  • Purpose: Integrates daily stock prices with annual statement disclosures to output daily historical multiples and statistical z-scores (standard deviations from the company's own historical average).
  • SQL Source: Derived from private transformation 9.5_valuation_multiples.sql (which ingests: prices and statements_usd from the staging schema; and stock_shares_outstanding from the raw schema).
  • Key Columns:
    • price_date (VARCHAR): Calendar date of the stock price.
    • pe_ratio (DOUBLE): Price / Trailing EPS.
    • price_to_fcf (DOUBLE): Market Cap / Free Cash Flow.
    • pe_z_score (DOUBLE): (Current PE - Mean historical PE) / StdDev(PE).
      • Interpretation: Negative z-scores (e.g. -1.5) indicate that the stock is historically cheap compared to its own historical trading range.

8. fct_macro_sensitivity

  • Purpose: Gauges how sensitive a stock's returns are to macroeconomic interest rate shifts by running a rolling 1-year linear regression of daily stock returns against daily changes in the 10-year US Treasury yield.
  • SQL Source: Derived from private transformation 9.6_macro_sensitivity.sql (which ingests: fct_macro_rates from the marts schema; and prices from the staging schema).
  • Key Columns:
    • treasury_beta_1y (DOUBLE): Slope coefficient of regression.
      • Interpretation: Positive beta (e.g. 1.2) means the stock moves up when bond yields rise (cyclical/financials). Negative beta (e.g. -0.8) means the stock falls when yields rise (utilities/defensives).
    • treasury_r2_1y (DOUBLE): R-squared (goodness of fit) of the rolling regression, showing how much of the stock's return variance is explained by yield changes.

9. fct_insider_sentiment

  • Purpose: Tracks buying and selling activity by corporate executives, officers, and directors.
  • SQL Source: Derived from private transformation 9.3_insider_sentiment.sql (which ingests: stock_insider_transactions from the raw schema).
  • Key Columns:
    • tx_date (DATE): Transaction execution date.
    • net_shares_daily (DOUBLE): Shares purchased minus shares sold on tx_date.
    • net_value_daily (DECIMAL): USD value of net transactions on tx_date.
    • rolling_30d_net_value / rolling_90d_net_value / rolling_180d_net_value (DECIMAL): Combined USD value of insider transactions over the respective trailing day windows.

10. fct_dividend_history

  • Purpose: Tracks distributions, stock splits, annual payouts, and consecutive dividend increase streaks.
  • SQL Source: Derived from private transformation 9.4_dividend_history.sql (which ingests: stock_dividend_events and stock_split_events from the raw schema).
  • Key Columns:
    • event_date (DATE): Date of dividend or split event.
    • event_type (VARCHAR): dividend or split.
    • amount (DOUBLE): Dividend distribution per share.
    • split_factor (VARCHAR): Split ratio (e.g. 2:1 or 1:1).
    • year_annual_payout (DECIMAL): Sum of all dividends paid in that calendar year.
    • year_dividend_increase_streak_years (BIGINT): The consecutive streak of years where the annual dividend payout increased.

11. fct_magic_formulas

  • Purpose: Implements Joel Greenblatt's "Magic Formula" screen, ranking companies by their return on capital and earnings yield.
  • SQL Source: Derived from private transformation 3_magic_formula.sql (which ingests: dim_company_profiles from the marts schema; prices and statements_usd from the staging schema; and stock_shares_outstanding and stock_valuation_snapshot from the raw schema).
  • Key Columns:
    • return_on_capital_pct (DOUBLE): EBIT / (Net Working Capital + Net PPE).
    • earnings_yield_pct (DOUBLE): EBIT / Enterprise Value.
    • roc_rank (BIGINT): Return on Capital percentile rank within the universe.
    • ey_rank (BIGINT): Earnings Yield percentile rank within the universe.
    • magic_formula_rank (BIGINT): Consolidated rank (sum of roc_rank and ey_rank).

12. fct_piotroski_fscores

  • Purpose: Computes Joseph Piotroski's 9-point binary score (0-9) analyzing profitability, leverage/liquidity, and operating efficiency.
  • SQL Source: Derived from private transformation 4_piotroski_fscore.sql (which ingests: statements_usd from the staging schema; and stock_shares_outstanding from the raw schema).
  • Key Columns:
    • f1_positive_roa to f9_improving_asset_turnover (INTEGER): Binary points (0 or 1) for each signal.
    • f_score (INTEGER): Combined health score (ranges from 0 to 9).
      • Interpretation: 8 or 9 is exceptionally strong; 0 to 3 is weak.
    • f_score_label (VARCHAR): Rating class (Strong Health, Moderate Health, Weak Health).

13. fct_altman_zscores

  • Purpose: Bankruptcy risk prediction using Edward Altman's 5-factor model for manufacturing and non-manufacturing firms.
  • SQL Source: Derived from private transformation 5_altman_zscore.sql (which ingests: prices and statements_usd from the staging schema; and stock_shares_outstanding and stock_valuation_snapshot from the raw schema).
  • Key Columns:
    • z_score (DOUBLE): Output score.
    • z_score_zone (VARCHAR): Solvency health zones:
      • Safe Zone (Z-Score > 2.90 for manufacturing, > 2.90 for service)
      • Grey Zone (1.23 <= Z-Score <= 2.90)
      • Distress Zone (Z-Score < 1.23, high risk of insolvency)

14. fct_shareholder_yields

  • Purpose: Computes cash returned to investors via dividends, stock buybacks, and net debt reduction.
  • SQL Source: Derived from private transformation 6_shareholder_yield.sql (which ingests: prices and statements_usd from the staging schema; and stock_shares_outstanding and stock_valuation_snapshot from the raw schema).
  • Key Columns:
    • dividend_yield_pct (DOUBLE): Cash dividends / Market Cap.
    • buyback_yield_pct (DOUBLE): Net stock buybacks / Market Cap.
    • debt_paydown_yield_pct (DOUBLE): Net debt paydown / Market Cap.
    • total_shareholder_yield_pct (DOUBLE): Combined yield of all three components.
    • tsy_label (VARCHAR): Rating based on shareholder yield strength.

15. fct_momentum_metrics

  • Purpose: Measures trend-following relative strength over multiple lookback windows (1m, 3m, 6m, 12m).
  • SQL Source: Derived from private transformation 7_momentum.sql (which ingests: prices from the staging schema).
  • Key Columns:
    • mom_12m_skip1m_pct (DOUBLE): Returns over 12 months excluding the most recent month (captures structural momentum while avoiding short-term reversal noise).
    • composite_momentum_pct (DOUBLE): Weighted average of 3m (20%), 6m (30%), and 12m-skip-1m (50%) momentum returns.
    • momentum_label (VARCHAR): Classification (Strong Momentum, Improving, Lagging, etc.).

16. fct_lynch_categories

  • Purpose: Implements Peter Lynch's stock categorization framework (Slow Grower, Stalwart, Fast Grower, Cyclical, Asset Play, Turnaround) using growth rates, leverage, size, and health filters.
  • SQL Source: Derived from private transformation 8_lynch_categories.sql (which ingests: dim_company_profiles, fct_altman_zscores, fct_financial_ratios, and fct_piotroski_fscores from the marts schema; and statements_usd from the staging schema).
  • Key Columns:
    • lynch_category (VARCHAR): The assigned Peter Lynch category.
    • lynch_confidence (VARCHAR): Strength classification of the assignment (High, Medium, Low).
    • net_cash_to_mktcap_pct (DOUBLE): Net balance sheet cash as a percentage of market cap (key for "Asset Play" screening).

17. fct_screener_recommendations

  • Purpose: A compound multi-factor grading model that weights Piotroski, Altman, Magic Formula, owner yield, growth, and R&D-adjusted efficiency into an overall long/short score.
  • SQL Source: Derived from private transformation 9_long_short_screener.sql (which ingests: dim_company_profiles, fct_altman_zscores, fct_financial_ratios, fct_magic_formulas, and fct_piotroski_fscores from the marts schema).
  • Key Columns:
    • long_score (INTEGER): Points accrued for high quality, value, health, and momentum (0 to 10 scale).
    • short_score (INTEGER): Points accrued for distress, manipulation risk, high debt, or poor cash conversion (0 to 10 scale).
    • long_short_recommendation (VARCHAR): Screener recommendation rating (e.g. STRONG_BUY when long_score >= 8 and short_score <= 1).

18. fct_capital_allocation

  • Purpose: Computes Return on Invested Capital (ROIC), pre-tax ROIC, and incremental ROIC over rolling 3-year and 5-year windows to evaluate management's capital deployment efficiency.
  • SQL Source: Derived from private transformation 9.18_capital_allocation.sql (which ingests: statements_usd from the staging schema).
  • Key Columns:
    • symbol (VARCHAR): Stock ticker symbol.
    • report_date (DATE): Statement report date.
    • tax_rate (DOUBLE): Effective tax rate (capped at 35%, default to 21%).
    • invested_capital (DOUBLE): Total Debt + Total Equity - Cash.
    • nopat (DOUBLE): Net Operating Profit After Tax (Operating Income * (1.0 - tax_rate)).
    • roic_pct (DOUBLE): Return on Invested Capital percentage.
    • pretax_roic_pct (DOUBLE): Pre-tax Return on Invested Capital percentage.
    • incremental_roic_3y_pct (DOUBLE): 3-year Incremental ROIC (using NOPAT / Invested Capital change).
    • pretax_incremental_roic_3y_pct (DOUBLE): 3-year Pre-tax Incremental ROIC (using EBIT / Invested Capital change).
    • incremental_roic_5y_pct (DOUBLE): 5-year Incremental ROIC.
    • pretax_incremental_roic_5y_pct (DOUBLE): 5-year Pre-tax Incremental ROIC.
    • capital_unlocked_growth_3y_flag (BOOLEAN): True if operating income grew while invested capital decreased over 3 years.
    • capital_allocation_category (VARCHAR): Classification of management's capital allocation efficiency (High-Efficiency Compounder, Fading Compounder, Turnaround Compounder, Value Destroyer, Capital-Light Grower, Standard Allocator).

19. fct_dcf_valuations

  • Purpose: Computes Weighted Average Cost of Capital (WACC), Cost of Equity (CAPM), Cost of Debt, and projects 10-year discounted cash flows under three growth scenarios (Base, Conservative, Aggressive).
  • SQL Source: Derived from private transformation 9.8_dcf_valuation.sql (which ingests: fct_financial_ratios, fct_growth_rates, and fct_macro_rates from the marts schema; statements_usd from the staging schema; and stock_valuation_snapshot from the raw schema).
  • Key Columns:
    • symbol (VARCHAR): Stock ticker symbol.
    • report_date (DATE): Reference report date for statement metrics.
    • wacc (DOUBLE): Weighted Average Cost of Capital (clamped between 5% and 15%).
    • cost_of_equity (DOUBLE): Cost of equity via CAPM (Risk-Free Rate + Beta * 5.5% ERP).
    • cost_of_debt (DOUBLE): Cost of debt (interest expense / total debt, or fallback).
    • tax_rate (DOUBLE): Effective tax rate.
    • base_cash_flow (DOUBLE): Owner earnings (falling back to FCF, then Net Income).
    • enterprise_value_base / enterprise_value_conservative / enterprise_value_aggressive (DOUBLE): Discounted enterprise values.
    • equity_value_base / equity_value_conservative / equity_value_aggressive (DOUBLE): Estimated equity value (Enterprise Value + Cash - Debt).
    • dcf_price_per_share_base / dcf_price_per_share_conservative / dcf_price_per_share_aggressive (DOUBLE): Estimated fair value per share.

20. fct_sector_benchmarks

  • Purpose: Computes peer-relative sector and industry percentiles and medians for multiple valuation, profitability, and momentum metrics.
  • SQL Source: Derived from private transformation 9.20_sector_benchmarks.sql (which ingests: dim_company_profiles, fct_capital_allocation, fct_financial_ratios, fct_growth_rates, and fct_momentum_metrics from the marts schema).
  • Key Columns:
    • symbol (VARCHAR): Stock ticker symbol.
    • sector (VARCHAR): Corporate macro sector.
    • industry (VARCHAR): Corporate micro industry.
    • pe_industry_percentile (DOUBLE): Industry percentile for P/E (0 = cheapest/best, 100 = most expensive/worst).
    • pb_industry_percentile (DOUBLE): Industry percentile for P/B.
    • roic_industry_percentile (DOUBLE): Industry percentile for ROIC (0 = highest/best, 100 = lowest/worst).
    • rev_growth_1y_industry_percentile (DOUBLE): Industry percentile for 1-year revenue growth.
    • mom_6m_industry_percentile (DOUBLE): Industry percentile for 6-month price momentum.
    • pe_sector_percentile (DOUBLE): Sector percentile for P/E.
    • mom_6m_sector_percentile (DOUBLE): Sector percentile for 6-month price momentum.
    • industry_median_pe / sector_median_pe (DOUBLE): Median P/E ratios.
    • industry_median_roic / sector_median_roic (DOUBLE): Median ROIC values.

21. fct_macro_rates

  • Purpose: Cleans and merges daily currency exchange rates and U.S. Treasury constant maturity yields.
  • SQL Source: Derived from private transformation 9.2_macro_rates.sql (which ingests: daily_treasury_yield and exchange_rate from the raw schema).
  • Key Columns:
    • currency_symbol (VARCHAR): Currency symbol (e.g. EURUSD=X).
    • report_date (DATE): Calendar date of the rates.
    • exchange_rate (DECIMAL): Daily close exchange rate.
    • bc_1month to bc_30year (DECIMAL): Constant maturity yields for 1m, 2m, 3m, 6m, 1y, 2y, 3y, 5y, 7y, 10y, 20y, and 30y U.S. Treasuries.

22. fct_earning_call_transcripts

  • Purpose: Compiles raw text paragraphs from quarterly earnings call transcripts and flags whether the speaker is a registered corporate officer (insider).
  • SQL Source: Derived from private transformation 9.21_fct_earning_call_transcripts.sql (which ingests: stock_earning_call_transcripts and stock_officers from the raw schema).
  • Key Columns:
    • symbol (VARCHAR): Stock ticker symbol.
    • fiscal_year (INTEGER): Fiscal year of the earnings call.
    • fiscal_quarter (INTEGER): Fiscal quarter.
    • paragraph_number (BIGINT): Section paragraph index.
    • speaker (VARCHAR): Name of the individual speaking.
    • is_insider (BOOLEAN): True if speaker matches a known corporate officer name for that symbol.
    • content (VARCHAR): Text content of the transcript paragraph.
    • transcripts_id (BIGINT): Unique identifier hash for the paragraph.
    • report_date (DATE): Earnings report date.

23. fct_sec_filing_sections

  • Purpose: Consolidates Item 7 (MD&A) and Item 1A (Risk Factors) sections from SEC corporate filings (e.g., 10-K, 10-Q) into a single textual database table.
  • SQL Source: Derived from private transformation 9.22_fct_sec_filing_sections.sql (which ingests: stock_sec_filing from the raw schema).
  • Key Columns:
    • symbol (VARCHAR): Stock ticker symbol.
    • accession_number (VARCHAR): Unique SEC accession identifier.
    • form_type (VARCHAR): Form type (e.g., 10-K or 10-Q).
    • filing_date (DATE): SEC filing submission date.
    • section_type (VARCHAR): Text category (mda or risk_factors).
    • section_text (VARCHAR): Full plaintext content extracted from the section.

Raw Schema Data Dictionary

This section describes the 25 raw ingestion tables loaded into the raw schema of database.db. These tables store primary data fetched from yFinance, SEC EDGAR, daily Treasury rate feeds, and exchange listings before downstream transformations are applied.

Table of Contents (Raw Schema)

  1. daily_treasury_yield (U.S. Treasury constant maturity yields)
  2. exchange_directories (Asset mapping & exchange lists)
  3. exchange_rate (Daily currency exchange rates)
  4. stock_analyst_price_targets (Sell-side consensus price targets)
  5. stock_analyst_recommendations (Buy/Sell recommendation matrices)
  6. stock_dividend_events (Chronological cash payouts)
  7. stock_earning_calendar (Earnings call dates and schedule)
  8. stock_earning_call_transcripts (Metadata for quarterly call transcript JSON files)
  9. stock_earnings_estimates (Analyst forward estimates)
  10. stock_earnings_history (Past quarterly EPS surprises)
  11. stock_eps_trends (Analyst consensus revisions)
  12. stock_insider_transactions (Form 4 executive and director trades)
  13. stock_institutional_holders (13F institutional ownership)
  14. stock_mutualfund_holders (Mutual fund holdings)
  15. stock_news (Aggregated financial news feeds)
  16. stock_officers (Company directors & executive compensation)
  17. stock_prices (Daily historical price bars)
  18. stock_profile (Sector, industry, and qualitative description)
  19. stock_revenue_breakdown (Segmented product/region revenues)
  20. stock_sec_filing (Filing metadata & raw text pointers)
  21. stock_shares_outstanding (Historical share counts)
  22. stock_split_events (Stock splits history)
  23. stock_statement (Faceted financial statements)
  24. stock_trailing_eps (Trailing earnings per share records)
  25. stock_valuation_snapshot (Latest yFinance snapshot metrics)

Raw Table Reference

daily_treasury_yield

  • Purpose: Daily constant maturity yields for U.S. government debt (from 1-month to 30-year bills/bonds).
  • Columns:
    • bc_1month to bc_30year (DECIMAL): Constant maturity yields for 1m, 2m, 3m, 6m, 1y, 2y, 3y, 5y, 7y, 10y, 20y, and 30y U.S. Treasuries.
    • report_date (DATE): Daily calendar date.

exchange_directories

  • Purpose: Asset mapping directories linking tickers, exchange details, asset type, and corporate country of origin.
  • Columns:
    • symbol (VARCHAR): Stock ticker symbol.
    • local_ticker (VARCHAR): Ticker symbol on local exchange.
    • exchange (VARCHAR): Ticker exchange code.
    • name (VARCHAR): Full company name.
    • asset_type (VARCHAR): Security type (e.g. stock, ETF).
    • country (VARCHAR): Country of corporate headquarters.

exchange_rate

  • Purpose: Daily close exchange rate value for currency cross-pairs.
  • Columns:
    • symbol (VARCHAR): Currency cross-pair symbol (e.g. EURUSD=X).
    • open / close / high / low (DECIMAL): Daily pricing values.
    • report_date (DATE): Calendar date.

stock_analyst_price_targets

  • Purpose: Consensus, low, high, mean, and median price targets estimated by sell-side analysts.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • current (DECIMAL): Current stock price.
    • low / high / mean / median (DECIMAL): Analyst price targets.
    • report_date (DATE): Extraction date.

stock_analyst_recommendations

  • Purpose: Aggregate recommendations matrix (number of analysts recommending strong buy, buy, hold, sell, strong sell) over various rolling periods.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • period (VARCHAR): Lookback period (e.g. 0m, -1m, etc.).
    • strong_buy / buy / hold / sell / strong_sell (INTEGER): Tally counts of recommendations.
    • report_date (DATE): Capture date.

stock_dividend_events

  • Purpose: Cash dividends declared and paid historically.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • amount (DECIMAL): Dividend payout amount per share.
    • report_date (DATE): Ex-dividend or payment date.

stock_earning_calendar

  • Purpose: Earnings announcement calendar, dates, and corresponding fiscal quarter info.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • time (VARCHAR): Before/after market close timing flag.
    • name (VARCHAR): Event description.
    • fiscal_quarter_ending (VARCHAR): Period end date.
    • report_date (DATE): Date of earnings release.

stock_earning_call_transcripts

  • Purpose: Metadata and directory path pointers for quarterly corporate earnings call transcript JSON text files.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • fiscal_year (INTEGER): Fiscal year of call.
    • fiscal_quarter (INTEGER): Fiscal quarter.
    • transcript_path (VARCHAR): File system location of raw JSON transcripts.
    • transcripts_id (INTEGER): Unique transcript identifier.
    • report_date (DATE): Reference period date.

stock_earnings_estimates

  • Purpose: Detailed forward consensus EPS/revenue estimate values, analyst counts, and target growth percentages.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • period (VARCHAR): Estimate target period.
    • estimate_type (VARCHAR): EPS or Revenue indicator.
    • avg_estimate / low_estimate / high_estimate (DECIMAL): Estimate stats.
    • number_of_analysts (INTEGER): Count of estimating analysts.
    • year_ago_value (DECIMAL): Historical matching period actual value.
    • growth (DECIMAL): Projected YoY growth rate.
    • currency (VARCHAR): Reporting currency.
    • report_date (DATE): Period reference date.

stock_earnings_history

  • Purpose: Tracks EPS surprise history by comparing actual quarterly EPS against consensus analyst estimates.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • quarter (VARCHAR): Target quarter.
    • eps_actual (DECIMAL): Realized EPS.
    • eps_estimate (DECIMAL): Expected EPS.
    • eps_difference (DECIMAL): Delta surprise value.
    • surprise_percent (DECIMAL): Surprise ratio.
    • report_date (DATE): Filing/calendar date.

stock_eps_trends

  • Purpose: Analyst EPS estimate revision trends showing revisions over 7, 30, 60, and 90-day horizons.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • period (VARCHAR): Forecast target period.
    • current_estimate (DECIMAL): Current average estimate.
    • days_7_ago / days_30_ago / days_60_ago / days_90_ago (DECIMAL): Historical estimates.
    • currency (VARCHAR): Invoiced currency.
    • report_date (DATE): As-of date.

stock_insider_transactions

  • Purpose: SEC Form 4 insider trading disclosures indicating trades executed by company officers and directors.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • insider (VARCHAR): Name of the corporate insider.
    • position (VARCHAR): Job title or relation to company.
    • transaction (VARCHAR): Transaction type (e.g. Sale, Buy, Option Exercise).
    • shares (BIGINT): Quantity of shares traded.
    • value (DECIMAL): Estimated transaction USD value.
    • ownership (VARCHAR): Direct or indirect ownership status.
    • url (VARCHAR): SEC Edgar filing URL.
    • text (VARCHAR): Brief transaction commentary.
    • report_date (DATE): Transaction filing date.
    • start_date (DATE): Trade execution date.

stock_institutional_holders

  • Purpose: Institutional ownership statistics based on SEC 13F filings.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • holder (VARCHAR): Institutional entity name.
    • pct_held (DECIMAL): Percentage of total shares outstanding owned.
    • shares (BIGINT): Share count.
    • value (BIGINT): Estimated USD value.
    • pct_change (DECIMAL): Change in shares held vs. prior filing.
    • report_date (DATE): Collection date.
    • date_reported (DATE): 13F filing reporting date.

stock_mutualfund_holders

  • Purpose: Mutual fund equity holder lists and ownership percentages.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • holder (VARCHAR): Mutual fund name.
    • pct_held / shares / value / pct_change (DECIMAL/BIGINT): Position sizes and changes.
    • report_date (DATE): Collection date.
    • date_reported (DATE): Report date.

stock_news

  • Purpose: Feeds of company-specific financial news articles and metadata.
  • Columns:
    • uuid (VARCHAR): Unique article ID.
    • symbol (VARCHAR): Associated ticker.
    • title (VARCHAR): Article headline.
    • publisher (VARCHAR): News source publisher.
    • report_date (DATE): Publication date.
    • type (VARCHAR): Category format.
    • link (VARCHAR): Web URL.
    • news (STRUCT): Nested structure containing paragraph details.
    • bucket_id (BIGINT): Storage grouping bucket.

stock_officers

  • Purpose: Directors, officers, key executives, salaries, and stock options details.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • name (VARCHAR): Officer name.
    • title (VARCHAR): Position title.
    • age (BIGINT): Executive's age.
    • born (BIGINT): Birth year.
    • pay (BIGINT): Total annual compensation in USD.
    • exercised / unexercised (BIGINT): Executed or outstanding options value.
    • report_date (DATE): Metadata capture date.

stock_prices

  • Purpose: Historical daily price bars (Open, Close, High, Low, Volume).
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • report_date (DATE): Price calendar date.
    • open / close / high / low (DECIMAL): Daily pricing indicators.
    • volume (BIGINT): Daily volume of shares traded.
    • bucket_id (BIGINT): Storage partition bucket.

stock_profile

  • Purpose: Qualitative company background, office address, industry classification, employee count, and website URL.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • address / city / country / phone / zip (VARCHAR): Corporate contact details.
    • industry / sector (VARCHAR): Industry and sector classifications.
    • long_business_summary (VARCHAR): Corporate business description.
    • full_time_employees (BIGINT): Count of employees.
    • web_site (VARCHAR): Corporate homepage URL.
    • report_date (DATE): Record capture date.

stock_revenue_breakdown

  • Purpose: Segmented corporate revenue breakdowns (by geographic region or business line).
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • breakdown (VARCHAR): Segment grouping category.
    • report_date (VARCHAR): Statement ending period date.
    • breakdown_name (VARCHAR): Segment name (e.g. North America, iPhone).
    • value (BIGINT): Revenue value.
    • period_type (VARCHAR): Period scale.
    • value_type (VARCHAR): Period or raw indicator.
    • series_name (VARCHAR): Statement series mapping.
    • currency (VARCHAR): Currency code.

stock_sec_filing

  • Purpose: Metadata and local text file system pointers for Item 7 MD&A and Item 1A Risk Factors from SEC filings.
  • Columns:
    • cik (VARCHAR): Central Index Key.
    • symbol (VARCHAR): Stock ticker.
    • company_name (VARCHAR): Corporate name.
    • form_type (VARCHAR): Form type (e.g. 10-K, 10-Q).
    • form_type_description (VARCHAR): SEC form description.
    • accession_number (VARCHAR): Unique SEC accession identifier.
    • acceptance_date_time (VARCHAR): System timestamp of submission acceptance.
    • filing_url (VARCHAR): Online filing path.
    • mda_text_path / risk_factors_text_path (VARCHAR): Path pointers to cleaned local text sections.
    • filing_date (DATE): Filing release date.
    • report_date (DATE): Reference period date.

stock_shares_outstanding

  • Purpose: Chronological corporate shares outstanding tracking history.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • shares_outstanding (BIGINT): Share count outstanding.
    • report_date (DATE): Reference period date.

stock_split_events

  • Purpose: Historic stock split coefficients and ratios.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • split_factor (VARCHAR): Split ratio (e.g. 2:1).
    • rn_1 (BIGINT): Row number sorting.
    • report_date (DATE): Effective split date.

stock_statement

  • Purpose: Normalized financial statement rows (balance sheet, income statement, cash flow) mapped to standard accounting items.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • item_name (VARCHAR): Standard statement item key.
    • item_value (DECIMAL): Accounting dollar amount.
    • finance_type (VARCHAR): Balance sheet, income statement, or cash flow indicator.
    • period_type (VARCHAR): annual or quarterly.
    • report_date (DATE): Statement ending period date.

stock_trailing_eps

  • Purpose: Trailing Twelve Months (TTM) earnings per share records.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • report_date (VARCHAR): Capture date.
    • trailing_eps (DECIMAL): Trailing EPS value.
    • update_time (VARCHAR): System timestamp of capture.

stock_valuation_snapshot

  • Purpose: Highly comprehensive daily metrics snapshot containing valuation ratios, growth rates, margin structures, balance sheet summaries, and trading metrics.
  • Columns:
    • symbol (VARCHAR): Stock ticker.
    • market_cap (BIGINT): Market capitalization.
    • trailing_pe / forward_pe / price_to_book / price_to_sales (DECIMAL): Core multiples.
    • enterprise_value (BIGINT): Corporate Enterprise Value.
    • enterprise_to_revenue / enterprise_to_ebitda (DECIMAL): Enterprise multiples.
    • beta (DECIMAL): Trading beta coefficient.
    • dividend_rate / dividend_yield (DECIMAL): Dividend summaries.
    • payout_ratio (DECIMAL): Dividend payout ratio.
    • ex_dividend_date (VARCHAR): Date of ex-dividend.
    • fifty_two_week_high / fifty_two_week_low (DECIMAL): Yearly price bounds.
    • fifty_day_average / two_hundred_day_average (DECIMAL): Moving price averages.
    • short_ratio / short_percent_of_float (DECIMAL): Short interest statistics.
    • return_on_assets / return_on_equity (DECIMAL): Asset and equity returns.
    • profit_margins / operating_margins (DECIMAL): Margin structures.
    • revenue_growth / earnings_growth (DECIMAL): Growth performance rates.
    • total_cash / total_debt (BIGINT): Debt and cash levels.
    • debt_to_equity / current_ratio / quick_ratio (DECIMAL): Leverage and liquidity ratios.
    • held_percent_insiders / held_percent_institutions (DECIMAL): Ownership concentration.
    • float_shares (BIGINT): Floating share count.
    • current_price (DECIMAL): Current close price.
    • exchange (VARCHAR): Trading exchange.
    • website (VARCHAR): Corporate homepage.
    • trailing_eps (DECIMAL): Trailing EPS.
    • report_date (DATE): Ingestion snapshot reference date.