1
2
3
Data Requirements
Upload File
Validation
Data Requirements
Production ML Models
CLIMA uses production-grade machine learning models with 0.842 default prediction AUC and 0.946 disaster prediction AUC (both exceed 0.80 industry standard). Validated on 145,452 observations and 68,485 FEMA disasters and 68,485 real disasters. The more data you provide, the more accurate our predictions will be.
To provide accurate CLIMA climate risk scores, we need loan and property data from your portfolio. Your data will be processed using our production ML models for superior risk assessment.
Required Fields
Loan ID
Unique identifier for each loan
Aliases: loan_number, loan_num, id, loan_identifier
Property State
2-letter state code (e.g., CA, FL, TX)
Aliases: state, prop_state, property_st
Priority Fields (Strongly Recommended) +0.04-0.05 AUC Boost
⭐ Flood Zone
Highest Impact
FEMA flood zone code (A, AE, V, VE, X, etc.)
You already have this from flood determination! Adds +0.04-0.05 AUC improvement to our production models (0.842/0.946 base AUC)
Aliases: fema_flood_zone, flood_zone_code, sfha, flood_determination
Property ZIP Code
5 or 9 digit ZIP code
Aliases: zip, zipcode, zip_code, postal_code
County FIPS
5-digit county FIPS code
Aliases: fips, county_code, fips_code
Current UPB
Current unpaid principal balance
Aliases: upb, unpaid_balance, current_balance, loan_balance
Original UPB
Original loan amount
Aliases: orig_upb, original_balance, loan_amount
Optional Fields (Improve Accuracy)
FICO Score
300-850
LTV Ratio
Loan-to-value
DTI Ratio
Debt-to-income
Property Type
SF, CONDO, PUD, etc.
Occupancy
Primary, Second, Investment
Property Value
Appraised value
Accepted File Formats
CSV
Excel (.xlsx, .xls)
Parquet
JSON
What You'll Get: Production-grade CLIMA scores with 0.842 default prediction AUC and 0.946 disaster prediction AUC, validated on 145,452 observations and 68,485 FEMA disasters. Includes Expected Loss (EL) calculations, PD/LGD adjustments, and portfolio-level analytics.
Download Template
Get a sample CSV template with all field names
Upload Your Portfolio Data
CSV, Excel, Parquet, or JSON files up to 500MB
Client Information
Validation Results
Total Records
Valid Records
Errors
Warnings
Field Coverage
Errors
... and more errors
File Validated Successfully!
Your data has been processed and is ready for CLIMA scoring using production ML models (0.842 default AUC, 0.946 disaster AUC).
Compare Two Loans
Compare candidate loans in the same area. Enter addresses and optional same-area ZIP for side-by-side CLIMA scores, risk tiers, and key drivers.
Loan A
Property score
Risk tier
Base (ZIP) score
Drivers:
Loan B
Property score
Risk tier
Base (ZIP) score
Drivers:
Same-area base: