Building Expert Scorecard

Example Scorecard: Demographic Risk Scorecard

This scorecard evaluates the risk of borrowers based on demographic factors like Age and Employment Status.

Step 1: Define the Categories

• Category 1: Demographic Data
  • Weight: 70%
• Category 2: Financial Stability
  • Weight: 30%
• Default Deficit Category: 0% (Total weight = 100%)

---

Step 2: Add Criteria to Each Category

Category: Demographic Data

• Criterion 1: Age
  • Type: Range
  • Weight: 50% of Demographic Data
• Criterion 2: Employment Status
  • Type: Variable
  • Weight: 50% of Demographic Data

Category: Financial Stability

• Criterion 1: Credit History
  • Type: Variable
  • Weight: 60% of Financial Stability
• Criterion 2: Debt-to-Income Ratio
  • Type: Range
  • Weight: 40% of Financial Stability

---

Step 3: Add Attributes (Bins) for Each Criterion

Criterion: Age (Range)

• Range and Impact:
  • 0–18: Impact = 20%
  • 18–30: Impact = 60%
  • 30–45: Impact = 100% (best range)
  • 45–65: Impact = 80%
  • 65–75: Impact = 40%

Criterion: Employment Status (Variable)

• Values and Impact:
  • Unemployed: Impact = 20%
  • Part-Time Employed: Impact = 50%
  • Fully Employed: Impact = 100% (best value)
  • Retired: Impact = 60%

Criterion: Credit History (Variable)

• Values and Impact:
  • No History: Impact = 10%
  • Poor History: Impact = 30%
  • Good History: Impact = 70%
  • Excellent History: Impact = 100% (best value)

Criterion: Debt-to-Income Ratio (Range)

• Range and Impact:
  • 0%–20%: Impact = 100% (best range)
  • 20%–40%: Impact = 80%
  • 40%–60%: Impact = 50%
  • 60%–80%: Impact = 30%
  • 80%–100%: Impact = 10%

---

Step 4: Assign Scorecard Points

• Scorecard Range: 0–1000 points.
• Category Points Distribution:
  • Demographic Data: 700 points (70% of total).
  • Financial Stability: 300 points (30% of total).
• Criteria Points Distribution (calculated based on category weights):
  • Age: 350 points (50% of 700).
  • Employment Status: 350 points (50% of 700).
  • Credit History: 180 points (60% of 300).
  • Debt-to-Income Ratio: 120 points (40% of 300).

---

Step 5: Fine-Tune and Adjust

• Use Scorport’s dynamic interface to:
  • Adjust weights for categories and criteria.
  • Refine the impacts of attributes (bins) to better reflect the data or expert judgment.

---

Step 6: Finalize and Save

• Once the scorecard is complete, it will Version 1.
• You can use this scorecard in the Scorport’s Strategy Designer to Generate as an Engine.

This model is done by the AI, using Scorport Expert Modeling Methodology.

---

Sample Outputs

Example Borrower:

• Age: 29 → 60% impact (210 points out of 350).
• Employment Status: Fully Employed → 100% impact (350 points out of 350).
• Credit History: Good History → 70% impact (126 points out of 180).
• Debt-to-Income Ratio: 50% → 50% impact (60 points out of 120).

Final Score:

210 + 350 + 126 + 60 = 746/1000

---

This Demographic Risk Scorecard is now ready to assess borrowers based on the defined weights and impacts.

import pandas as pd

# Define the structure of the scorecard as a dictionary
scorecard_data = {
    "Category": ["Demographic Data", "Demographic Data", "Financial Stability", "Financial Stability"],
    "Criterion": ["Age", "Employment Status", "Credit History", "Debt-to-Income Ratio"],
    "Type": ["Range", "Variable", "Variable", "Range"],
    "Weight in Category (%)": [50, 50, 60, 40],
    "Points Allocated": [350, 350, 180, 120],
    "Attributes / Bins": [
        "0–18: 20%, 18–30: 60%, 30–45: 100%, 45–65: 80%, 65–75: 40%",
        "Unemployed: 20%, Part-Time Employed: 50%, Fully Employed: 100%, Retired: 60%",
        "No History: 10%, Poor History: 30%, Good History: 70%, Excellent History: 100%",
        "0–20%: 100%, 21–40%: 80%, 41–60%: 50%, 61–80%: 30%, 81–100%: 10%",
    ],
}

# Convert to a DataFrame
scorecard_df = pd.DataFrame(scorecard_data)

# Display the table
import ace_tools as tools; tools.display_dataframe_to_user(name="Scorport Expert Scorecard", dataframe=scorecard_df)

This model code is done by the AI, using Scorport Expert Modeling Methodology.