[Feature] Translate raw Alt-data into actionable investment stratergies

[Darshan174]

Problem

Currently, AlphaRadar successfully detects and surfaces alternative data anomalies (hiring surges, executive moves) and generates a FusedInsight score. While valuable, this leaves the user asking "So what should I buy/short?". The platform needs to bridge the gap between interesting data and actionable portfolio execution.

Proposed Solution

Enhance the backend fusion.py / scorer.py engines and the frontend UI to explicitly suggest trade setups and risk profiles based on the detected signals.

Key Tasks

1. Signal Historical Backtesting (The "Prove It" Layer)

• Calculate and attach historical win-rates to signals (e.g., "NVDA hiring surges historically precede a +8% 3-month return").
• Update UI to explicitly show the statistical edge of each active signal tag.

2. Algorithmic "Pairs Trade" Generation

• Modify sector_pulse and competitor detectors to actively search for extreme divergences between rival companies.
• Generate market-neutral "Pairs Trade" alerts in the UI (e.g., "Long AMD / Short INTC based on 30-day talent migration").

3. Earnings-Catalyst Cross-Referencing

• Ingest upcoming financial calendar/earnings dates.
• Create a "Pre-Earnings Setup" view linking extreme alt-data FusedInsights to companies reporting within 14 days.

4. Advanced Risk & Short-Selling Flags

• Explicitly tag highly negative combined signals (e.g., exec exodus + hiring freeze while stock ATH) as "Take Profit" or "Short Candidate" alerts rather than just low scores.

5. LLM Time Horizon Classification

• Update the Groq/Llama prompt in analysis/ to output a structured time horizon for the anomaly (e.g., Short-term Catalyst, Long-term Compounder, Value Trap).