Skip to content

DBN Analytics

Data Model

dbn-analytics-poc

Data Model¶

The SQLite database (dbn-poc-database.db) contains a single table, msmeloans, with ~11,000 rows and 40 columns.

Table: `msmeloans`¶

DDL¶

CREATE TABLE "msmeloans" (
  "SeqNo"                          REAL,
  "Gender"                         TEXT,
  "State"                          TEXT,
  "Sector"                         TEXT,
  "PFI ID"                         TEXT,
  "FullNames"                      TEXT,
  "BirthDate"                      TEXT,
  "AmountGranted"                  INTEGER,
  "EffectiveDate"                  TEXT,
  "Tenor"                          INTEGER,
  "TenorType"                      TEXT,
  "MaturityDate"                   TEXT,
  "TenorInDays"                    INTEGER,
  "Rate"                           REAL,
  "ScheduleType"                   TEXT,
  "InterestRepayStartDate"         TEXT,
  "PrincipalRepayStartDate"        TEXT,
  "InterestRepayFreq"              TEXT,
  "PrincipalRepayFreq"             TEXT,
  "NatureOfBusiness"               TEXT,
  "FirstTimeAccessToCredit"        TEXT,
  "StartUp"                        TEXT,
  "MSMEAnnualTurnover"             INTEGER,
  "NumberOfEmployees"              INTEGER,
  "Age"                            INTEGER,
  "age_group_id"                   INTEGER,
  "Status"                         INTEGER,
  "DateAdded"                      TEXT,
  "ES_Rating"                      TEXT,
  "green_energy/energy_efficiency" TEXT,
  "FirstTimeAccessToCredit_norm"   INTEGER,
  "StartUp_norm"                   INTEGER,
  "age_group"                      TEXT,
  "loan_to_turnover"               REAL,
  "employees_bucket"               TEXT,
  "synthetic_default"              INTEGER,
  "risk_score"                     REAL,
  "risk_category"                  TEXT,
  "pred_default_prob"              REAL,
  "predicted_default"              INTEGER
);

Column Reference¶

Borrower Demographics¶

Column	Type	Description
`SeqNo`	REAL	Sequential record number
`Gender`	TEXT	`M` or `F`
`State`	TEXT	State ID of the borrower
`FullNames`	TEXT	Full name of the borrower
`BirthDate`	TEXT	Date of birth (`YYYY-MM-DD`)
`Age`	INTEGER	Age in years
`age_group`	TEXT	Age bracket (e.g., `36-45`)
`age_group_id`	INTEGER	Numeric ID for age bracket

Loan Details¶

Column	Type	Description
`PFI ID`	TEXT	Participating Financial Institution identifier
`AmountGranted`	INTEGER	Loan amount in local currency
`EffectiveDate`	TEXT	Date the loan was disbursed
`MaturityDate`	TEXT	Date the loan matures
`Tenor`	INTEGER	Loan duration
`TenorType`	TEXT	Unit of tenor (e.g., `Months`)
`TenorInDays`	INTEGER	Tenor expressed in days
`Rate`	REAL	Interest rate (%)
`ScheduleType`	TEXT	Repayment schedule type
`InterestRepayStartDate`	TEXT	First interest repayment date
`PrincipalRepayStartDate`	TEXT	First principal repayment date
`InterestRepayFreq`	TEXT	Frequency of interest repayment
`PrincipalRepayFreq`	TEXT	Frequency of principal repayment
`Status`	INTEGER	Loan status code
`DateAdded`	TEXT	Record creation date

Business Profile¶

Column	Type	Description
`Sector`	TEXT	Business sector (e.g., `Trade and Commerce`)
`NatureOfBusiness`	TEXT	Nature of business description
`MSMEAnnualTurnover`	INTEGER	Annual business turnover
`NumberOfEmployees`	INTEGER	Number of employees
`employees_bucket`	TEXT	Categorised employee band (e.g., `Solo`, `Micro`)
`loan_to_turnover`	REAL	Ratio of loan to annual turnover
`FirstTimeAccessToCredit`	TEXT	Whether this is the first credit access
`FirstTimeAccessToCredit_norm`	INTEGER	Normalised: `1` = first time, `0` = returning
`StartUp`	TEXT	Whether the business is a startup
`StartUp_norm`	INTEGER	Normalised: `1` = startup, `0` = established
`ES_Rating`	TEXT	Environmental & Social rating
`green_energy/energy_efficiency`	TEXT	Green energy flag

Risk & Prediction¶

Column	Type	Description
`synthetic_default`	INTEGER	Simulated default flag (`0` or `1`)
`risk_score`	REAL	Composite risk score (`0.0–1.0`)
`risk_category`	TEXT	Risk label (e.g., `Low Risk`, `High Risk`)
`pred_default_prob`	REAL	ML-predicted probability of default (`0.0–1.0`)
`predicted_default`	INTEGER	Binary prediction: `1` = will default

Sample Data¶

SELECT FullNames, Sector, AmountGranted, pred_default_prob, risk_category
FROM msmeloans
ORDER BY pred_default_prob DESC
LIMIT 5;

FullNames	Sector	AmountGranted	pred_default_prob	risk_category
James Okafor	Agriculture	8,000,000	0.92	High Risk
Aisha Bello	Manufacturing	5,500,000	0.87	High Risk
...	...	...	...	...