Train Predictive Model

Training Requirements:

To train the predictive model, you need:

  • Multiple PDF Transcripts - At least 3-5 earnings call transcripts with proper naming convention
  • Excel File with Earnings Data - Historical performance data matching the transcripts
1. Upload Transcript Files

Naming Convention: [TICKER]4Q24-Earnings-Call-Transcript.pdf
Examples: CPB4Q24-Earnings-Call-Transcript.pdf or 4Q24-Earnings-Call-Transcript.pdf

Drag & Drop PDF Transcripts

or click to browse

2. Upload Earnings Data
Upload Excel File

with earnings performance data

Excel File Format Requirements:
Required Columns:
  • Column 1: Earnings Quarter using format 4Q24, 1Q25 (must match transcript filename)
  • Last Column: Stock Price Movement (% change, e.g., "+5.2", "-3.1")
Optional Columns (between Quarter and Stock Movement):
  • EPS vs Expectations: "Beat", "Miss", or "Meet"
  • Guidance vs Expectations: "Beat", "Miss", or "Meet"
Example: Earnings | EPS_vs_Expectations | Guidance_vs_Expectations | Stock_Reaction
4Q24 | Beat | Meet | +4.5
What the model learns:
Sentiment Patterns

How emotional tone correlates with stock movement

Call Strength

Confidence levels and financial strength indicators

Concern Signals

Evasive language and uncertainty patterns

Performance Data

EPS and guidance vs expectations impact