MAG 7 PRICES bundle 2020-2025 - Daily AI Feature Feed

What You Get

• 10,556 trading days (1,508 per ticker x 7 tickers)
• 95 columns (ticker identifier + 3 metadata + 84 features + 7 forward labels)
• Single CSV file (~7.4 MB), sorted by ticker then date
• Filter by ticker column for single-stock or cross-stock analysis
• One row per ticker per trading day — clean, ready for ML ingestion

Bundle Savings

Feature Groups (95 columns)

Why the Magnificent 7 Bundle

• Cross-stock sentiment divergence — when NVDA sentiment turns bullish while AAPL turns bearish, it signals sector rotation, not market direction. Individual datasets cannot capture this.
• Pairs trading signals — sentiment spread between correlated pairs (MSFT/GOOG, NVDA/AMD-exposed tickers, AMZN/MSFT cloud) provides mean-reversion and momentum signals
• Portfolio-level regime detection — when all 7 stocks share the same sentiment direction, it signals macro regime change. When they diverge, it signals idiosyncratic opportunity.
• Relative value modeling — 84 features x 7 tickers = 588 feature comparisons per day for relative strength models
• Contagion analysis — track how sentiment shocks in one mega-cap (e.g., NVDA earnings miss) propagate through the other 6 over subsequent days

Collection Methodology

1. News Collection: GDELT (Global Database of Events, Language, and Tone) provides 100+ articles daily per ticker from global financial media
2. AI Sentiment Analysis: Google Gemini processes each article batch per ticker independently, scoring overall market sentiment and ticker-specific sentiment on a -1.0 to +1.0 scale
3. Price Data: Yahoo Finance provides closing prices for all 7 tickers, S&P 100 cross-asset data, sector returns, VIX, yields, credit spreads, and options data
4. Feature Engineering: 94 features per ticker are computed from raw inputs using numpy — technical indicators, sentiment derivatives, cross-asset correlations, macro signals, and forward labels. A ticker identifier column is prepended for the bundle.
5. Quality Control: Automated audit checks coverage, NaN rates, column integrity, date alignment across all 7 tickers, and data consistency

Complete Column Reference

# | Column | Type | Group | Description
---|--------|------|-------|------------
1 | ticker | string | Identifier | Stock symbol (NVDA, AAPL, TSLA, AMZN, META, MSFT, GOOG)
2 | date | date | Metadata | Trading date (YYYY-MM-DD)
3 | year | int | Metadata | Calendar year
4 | day_of_week | string | Metadata | Day name (Monday-Friday)
5 | price_close | float | Price Action | Closing price (USD)
6 | price_return_1d | float | Price Action | 1-day return (%)
7 | price_return_3d | float | Price Action | 3-day return (%)
8 | price_return_5d | float | Price Action | 5-day return (%)
9 | price_return_10d | float | Price Action | 10-day return (%)
10 | price_return_20d | float | Price Action | 20-day return (%)
11 | price_log_return_1d | float | Price Action | 1-day log return
12 | price_dist_high_10d | float | Price Action | Distance from 10-day high (%)
13 | price_dist_high_20d | float | Price Action | Distance from 20-day high (%)
14 | price_dist_low_10d | float | Price Action | Distance from 10-day low (%)
15 | price_dist_low_20d | float | Price Action | Distance from 20-day low (%)
16 | tech_rsi_14 | float | Technical | RSI 14-day (0-100, >70 overbought, <30 oversold)
17 | tech_sma_5 | float | Technical | 5-day Simple Moving Average (USD)
18 | tech_sma_20 | float | Technical | 20-day Simple Moving Average (USD)
19 | tech_sma_5_dist | float | Technical | Distance from 5-day SMA (%)
20 | tech_sma_20_dist | float | Technical | Distance from 20-day SMA (%)
21 | tech_ema_20 | float | Technical | 20-day Exponential Moving Average (USD)
22 | tech_ema_20_dist | float | Technical | Distance from 20-day EMA (%)
23 | tech_bollinger_pos | float | Technical | Position within Bollinger Bands (0=lower, 0.5=middle, 1=upper)
24 | tech_bollinger_bw | float | Technical | Bollinger Band width (%)
25 | tech_bollinger_squeeze | float | Technical | Squeeze indicator (1 = bandwidth in bottom 10th percentile, upcoming breakout)
26 | tech_macd | float | Technical | MACD line
27 | tech_macd_signal | float | Technical | MACD signal line
28 | tech_macd_hist | float | Technical | MACD histogram (positive = bullish momentum)
29 | tech_adx | float | Technical | ADX trend strength (>25 trending, <20 ranging)
30 | tech_roc_5 | float | Technical | 5-day Rate of Change (%)
31 | tech_roc_10 | float | Technical | 10-day Rate of Change (%)
32 | tech_roc_20 | float | Technical | 20-day Rate of Change (%)
33 | tech_streak | int | Technical | Consecutive up/down days (positive = up streak)
34 | tech_vol_5d | float | Technical | 5-day realized volatility (annualized %)
35 | tech_vol_10d | float | Technical | 10-day realized volatility (%)
36 | tech_vol_20d | float | Technical | 20-day realized volatility (%)
37 | tech_momentum_quality | float | Technical | Momentum consistency score (ROC adjusted for volatility)
38 | sent_overall | float | AI Sentiment | Overall market sentiment (-1.0 bearish to +1.0 bullish)
39 | sent_{ticker} | float | AI Sentiment | Ticker-specific sentiment (-1.0 bearish to +1.0 bullish)
40 | sent_spread | float | AI Sentiment | Ticker minus overall sentiment (positive = stock more bullish than market)
41 | sent_overall_ma3 | float | AI Sentiment | 3-day MA of overall sentiment
42 | sent_{ticker}_ma3 | float | AI Sentiment | 3-day MA of ticker sentiment
43 | sent_overall_ma5 | float | AI Sentiment | 5-day MA of overall sentiment
44 | sent_{ticker}_ma5 | float | AI Sentiment | 5-day MA of ticker sentiment
45 | sent_overall_mom3 | float | AI Sentiment | 3-day overall sentiment momentum
46 | sent_{ticker}_mom3 | float | AI Sentiment | 3-day ticker sentiment momentum
47 | sent_overall_mom5 | float | AI Sentiment | 5-day overall sentiment momentum
48 | sent_{ticker}_mom5 | float | AI Sentiment | 5-day ticker sentiment momentum
49 | sent_overall_vol5 | float | AI Sentiment | 5-day overall sentiment volatility
50 | sent_overall_vol10 | float | AI Sentiment | 10-day overall sentiment volatility
51 | sent_{ticker}_vol5 | float | AI Sentiment | 5-day ticker sentiment volatility
52 | sent_price_corr_10d | float | Sent x Price | 10-day rolling sentiment-price correlation
53 | sent_price_corr_20d | float | Sent x Price | 20-day rolling sentiment-price correlation
54 | sent_price_diverge | float | Sent x Price | Divergence flag (1 when sentiment and price disagree 3+ consecutive days)
55 | news_count | int | News Volume | Articles collected that day
56 | news_count_ma20 | float | News Volume | 20-day article count moving average
57 | news_count_zscore | float | News Volume | Z-score vs 20-day window (>2 = unusual volume)
58 | news_count_mom5 | float | News Volume | 5-day article count momentum
59 | news_spike | float | News Volume | Binary flag for abnormal news volume (count > 2x MA)
60 | mkt_sp100_breadth | float | Cross-Asset | S&P 100 market breadth (% of stocks up, 0-100)
61 | mkt_sp100_return | float | Cross-Asset | S&P 100 equal-weight return (%)
62 | mkt_dispersion | float | Cross-Asset | Cross-sectional return dispersion (%)
63 | mkt_tech_return | float | Cross-Asset | Tech sector return (%)
64 | mkt_finance_return | float | Cross-Asset | Finance sector return (%)
65 | mkt_health_return | float | Cross-Asset | Healthcare sector return (%)
66 | mkt_energy_return | float | Cross-Asset | Energy sector return (%)
67 | mkt_consumer_return | float | Cross-Asset | Consumer sector return (%)
68 | mkt_industrial_return | float | Cross-Asset | Industrial sector return (%)
69 | mkt_{ticker}_vs_tech | float | Cross-Asset | Ticker minus tech sector return (%)
70 | mkt_{ticker}_vs_market | float | Cross-Asset | Ticker minus S&P 100 return (%)
71 | mkt_{ticker}_beta_20d | float | Cross-Asset | 20-day rolling beta vs S&P 100
72 | vol_of_vol_20d | float | Vol Regime | Volatility of volatility (20-day rolling)
73 | vol_regime | int | Vol Regime | Regime: 0=low, 1=normal, 2=high, 3=extreme
74 | sent_vol_regime_interaction | float | Interaction | Ticker sentiment x volatility regime (amplified signal in high-vol periods)
75 | macro_vix | float | Macro | VIX level (CBOE Volatility Index)
76 | macro_vix_change_1d | float | Macro | 1-day VIX percentage change
77 | macro_vix_ma5 | float | Macro | VIX 5-day moving average
78 | macro_yield_spread | float | Macro | 10Y minus short-term Treasury yield spread (%)
79 | macro_credit_spread | float | Macro | High-yield credit spread proxy (LQD/HYG ratio)
80 | macro_dxy_change | float | Macro | US Dollar Index 1-day percentage change (%)
81 | options_iv_atm | float | Options | ATM implied volatility (%) via VIX x beta approximation
82 | options_iv_rv_spread | float | Options | IV minus realized vol (positive = fear premium)
83 | earnings_days_to_next | float | Earnings | Days to next quarterly earnings (0 = earnings day, capped at 90)
84 | earnings_is_earnings_day | float | Earnings | Binary flag: 1 on earnings day, 0 otherwise
85 | attention_wikipedia_views | float | Social Attention | Daily Wikipedia pageviews (1-day lagged)
86 | attention_wikipedia_zscore | float | Social Attention | Wikipedia views z-score (>2 = unusual interest)
87 | attention_google_trends | float | Social Attention | Google Trends index (0-100)
88 | attention_google_trends_change | float | Social Attention | Google Trends change vs prior period
89 | label_return_1d | float | Forward Labels | Next-day return (%)
90 | label_dir_1d | string | Forward Labels | Next-day direction (UP/DOWN/FLAT)
91 | label_return_3d | float | Forward Labels | 3-day forward return (%)
92 | label_dir_3d | string | Forward Labels | 3-day forward direction
93 | label_return_5d | float | Forward Labels | 5-day forward return (%)
94 | label_dir_5d | string | Forward Labels | 5-day forward direction
95 | label_flat_1d | float | Forward Labels | Flat flag (1 if next-day |return| < 0.1%)

Data Quality

• NaN values limited to first ~30 rows per ticker (indicator warmup period)
• Last 1-5 rows per ticker may have empty forward labels (not yet realized)
• Zero gaps in sentiment coverage — every trading day has article data for every ticker
• All 7 tickers share identical date ranges (1,508 trading days each)
• Quality rating: 5/5 (automated audit verified)

Known Limitations

• Ticker-specific sentiment columns (e.g., sent_nvda) are derived from AI analysis of English-language financial news via GDELT. Non-English sources and social media are not included.
• Options IV (options_iv_atm) is approximated using VIX x ticker beta, not from actual options chains. This standard quant approximation may diverge from true ATM IV during earnings or extreme moves.
• Cross-asset features require both the ticker and the reference asset to trade on the same day. Holidays in one market produce NaN for that pair.
• Earnings dates are pattern-estimated with yfinance confirmation. Accuracy is +/- 3 days for some historical quarters.
• The flat zone threshold in forward labels is 0.1% for the bundle (tighter than individual datasets) to capture meaningful cross-stock moves.

Use Cases

• Cross-stock ML models — predict relative performance across the Magnificent 7 using 95 daily features
• Pairs trading — exploit sentiment divergence between correlated mega-caps (e.g., MSFT vs GOOG, NVDA vs AMZN cloud capex)
• Portfolio construction — build Mag 7 allocation models using sentiment, technical, and macro signals
• Sector rotation — detect regime shifts across AI, cloud, EV, and consumer tech
• Contagion research — study how sentiment shocks propagate across mega-caps
• Academic research — 6 years x 7 stocks of sentiment-price dynamics across multiple market regimes
• LLM fine-tuning — train financial reasoning models on structured cross-stock market data

Pairs Well With

License

AI Training Rights