Senti.

MSc Data Science Research

Every senti counts.

Senti is a research project exploring machine learning-based credit scoring for informal waste sector workers. Built to prove that transaction data can unlock financial inclusion for those without formal credit histories.

About Senti

Senti is a research prototype developed as part of an MSc Data Science thesis at the University of Dar es Salaam. The project investigates how operational data from waste collection centres can be transformed into creditworthiness indicators for informal workers.

We started with a simple question: why can't a waste picker get a loan? The answer wasn't lack of income—it was lack of records. No transaction history. No way to prove earnings. No financial identity.

Senti addresses this by digitizing collection centre transactions—every weighment, every payment, every worker—and using machine learning to derive credit scores from this operational data.

Research Focus

This project combines data engineering, machine learning, and financial inclusion research to develop alternative credit scoring models for underserved populations.

Data Collection

Digital tools for capturing transaction data at collection centres—weighments, payments, worker identification—replacing paper ledgers with structured datasets.

Feature Engineering

Extracting creditworthiness signals from raw transaction data—consistency patterns, volume trends, relationship tenure, and behavioural indicators.

ML Credit Scoring

Developing and validating machine learning models that predict creditworthiness without traditional credit bureau data or formal employment records.

Financial Inclusion

Demonstrating how alternative data sources can bridge the gap between informal sector workers and formal financial services.

Why "Senti"

In Swahili, "senti" means cent—the smallest denomination of currency. When people say "senti senti," they mean something insignificant. Too small to matter.

We named this project Senti because we believe the opposite. The waste economy runs on small transactions—a few shillings per kilogram, a handful of bottles at a time.

This research aims to make every senti visible. To prove that small, consistent work adds up to creditworthiness. To give waste workers the financial recognition they've earned.

Kila senti inahesabika. Every cent counts.

Research Data

4,000+
Waste pickers in dataset
12
Collection centres
Tanga
Study area

Data collected in partnership with Zaidi Recyclers Limited.

University of Dar es Salaam ¡ MSc Data Science

Contact

For inquiries about this research or collaboration opportunities, please reach out.

Get in Touch