String Analytics develops custom credit scoring engines that can create credit scores based on a set of underlying data provided by different institutions and different data sources.
The scoring engine runs the credit scoring on a real- time / daily / monthly basis depending on the frequency of receipt of the underlying data. The models deployed are based on Machine Learning algorithms
data processing
The platform is able to process large datasets from multiple data sources
Alternative datasets.
Msisdn
Topup amounts
Topup frequencies
Device types
SIM activation dates
Location data
Minutes of use (Onnet & Offnet)
Spend on network (sms,airtime,data,bundles)
Last revenue generated event.
Agricultural input data
Transactional datasets.
Loan history
Average repayment time
Credit score
Repayment amounts
No of successful repayments
Conventional datasets.
Cash deposit/withdrawal amounts
Cash deposit /withdrawal frequencies
Bill payment info
Registration details on mobile wallet/ Bank account
Wallet to wallet (account to account) transferdetails