S&P 500 swing dashboard — AI pattern detection, valuation screening, and after-close signals.
This tool is for research and education only. Nothing shown here constitutes financial advice, a recommendation to buy or sell, or a guarantee of any outcome. Signals can be wrong, delayed, or incomplete. You are solely responsible for your trades, risk management, and any financial losses. Always do your own research and consult a licensed financial adviser before making investment decisions.
This dashboard has four tabs: Guide (this page), Opportunities (PEG) for screening the S&P 500 by valuation, Chart + Levels for per-ticker technical analysis, and Daily Report for an after-close summary of signals.
Typical workflow: check the Daily Report tab each evening after the data refresh → use 52W extremes and pattern signals to build a watchlist → scan the Opportunities tab for valuation context → click a row to jump to that ticker's chart → study levels, gaps, and patterns to form a thesis.
The table shows all S&P 500 tickers ranked by valuation. Key columns:
• Discount %: how far the current price is below the analyst target (positive = upside, negative = already above target). E.g. AAPL at 15% discount means the average analyst target is 15% above the current price.
• PEG: Price/Earnings-to-Growth ratio. A PEG below 1.0 is often considered "growth at a reasonable price". E.g. PEG = 0.8 on a fast grower is more attractive than PEG = 3.0 on a slow grower.
• Undervalued: discount ≥ 10% AND PEG ≤ 2. Deep Value: stricter — discount ≥ 20% AND PEG ≤ 1.5. Overvalued: price is already above target by more than 10%.
Click any row to open that ticker's chart automatically.
Support is a price zone where buyers have previously stepped in, causing price to bounce up. Resistance is a zone where sellers have pushed price back down.
Example: if SPY bounced off $450 three times in the past, $450 is strong support. If it has been rejected at $470 twice, $470 is resistance.
Primary: highest-confidence level (most recent touches, multiple hits). Secondary: next best level. Fallback: auto-generated anchor when no strong detected level exists on one side.
A bull gap occurs when today's open is above yesterday's high — leaving an unfilled region below. A bear gap is the reverse. Markets often "fill" gaps later.
Example: NVDA gaps up 5% on earnings — the gap low and high are marked. If price dips back to that range, it is testing the gap fill zone.
Heuristic best-fit lines drawn through recent swing pivots. A rising support trendline connecting higher lows suggests an uptrend. Toggle with the Show Trendlines checkbox.
These are candidate lines only — not guaranteed to hold. Use them as context alongside levels.
Geometric pattern detection (flags, wedges, triangles, double tops/bottoms) from recent price bars. Toggle with Show Patterns.
Example: a bull flag is a sharp up-move followed by a tight, slightly downward channel — often resolves to the upside. A head and shoulders is a reversal pattern with three peaks, the middle being highest.
A YOLOv8 computer-vision model (foduucom/stockmarket-pattern-detection-yolov8) is run nightly on each ticker's candlestick chart image. Detected patterns are stored and shown as dotted bounding boxes on the chart.
Pattern types: Head & Shoulders Top/Bottom, M-Head, W-Bottom, Triangle, StockLine. Confidence is shown as a percentage — higher is more certain.
Toggle with YOLO Patterns checkbox. These are updated each trading day and do not require any waiting on your end.
Combines rule-based geometry confidence with candlestick alignment, volume regime, and breakout state into a single score. Higher hybrid confidence means more corroborating evidence.
An image-style shape scorer using normalised OHLC geometry — not a trained deep model, but useful as a secondary signal. Compare with Hybrid scores for consensus.
Generated automatically after each daily data refresh (22:00 UTC Mon–Fri). The weekly YOLO pass runs separately on Saturdays and also triggers a report update. The report does not have a manual refresh button — it reflects the last completed cron run.
52-Week Extremes — lists every S&P 500 ticker whose latest close is within 3% of its 52-week high or low. Click any ticker to jump directly to its chart.
Rationale: proximity to a 52W high is a momentum signal — stocks making new highs are often in strong uptrends and may continue. Proximity to a 52W low flags potential capitulation or value traps. In both cases, the price is at a structurally significant level and worth further inspection.
A ticker at its 52W high is not automatically bullish — check volume (a breakout on low volume is weak) and whether the high is a genuine new multi-year high or just a recovery within a longer downtrend. Similarly, a 52W low could be genuine distress or a brief oversold condition before recovery.
Top AI Patterns at Close (YOLO) — the 30 highest-confidence chart patterns detected by the YOLOv8 model on the latest trading day's data. Sorted by confidence. The timeframe column shows whether the detection came from the 180-day daily chart or the 730-day weekly chart.
Rationale: rather than manually checking all 510 tickers, this surfaces the detections the model is most confident about. Patterns detected on the weekly timeframe represent larger, more structurally significant formations.
Candle Signals at Close — rule-based candlestick patterns detected on each ticker's latest closing candle. Common signals include engulfing candles, hammers, shooting stars, and three-candle reversal sequences (morning star, evening star, three white soldiers, three black crows).
Rationale: candlestick signals are context-dependent — a hammer after a prolonged downtrend on high volume is far more meaningful than a hammer in the middle of a range. Use this list as a starting point to check the chart, not a standalone trade signal. Bullish and bearish signals are both shown; filter by bias to focus on one direction.
The 3M / 6M / YTD / 1Y / 2Y / ALL buttons at the top-right of the chart change the view window. All data is always loaded — changing the window just zooms in. You can also drag to zoom, double-click to reset, and scroll to zoom.
The engine uses a wider history window (up to all available data) for more stable level and pivot calculations, regardless of what is shown on screen.
• Pattern overfitting: a pattern that worked in backtests may not work going forward.
• Survivorship bias: S&P 500 constituents are already winners — they have survived long enough to be listed here.
• Regime shifts: patterns and levels from a bull market may not behave the same in a bear market or high-volatility regime.
• Low-liquidity distortions: thin stocks can have false breakouts. Always check volume.
Use strict position sizing and stop-losses. Never risk more than you can afford to lose.