INSCHOOL
by Open University of Tanzania
A lightweight, privacy‑aware tool that helps schools log attendance and use Machine Learning (ML) to identify students at risk—so educators can act early.
Built by The Open University of Tanzania in collaboration with AiGI Systems Company Limited. Designed for low‑resource settings. Backed by evidence from pilots across ~70 districts.
INSCHOOL is an open‑source ML powered digital attendance tool plus an early‑warning model that flags students at risk of dropout based on school‑level signals. It is optimized for offline‑first usage with periodic sync when connectivity is available.
• Schools: simple daily attendance and risk lists.
•
Researchers: transparent, reproducible models.
• Developers: modular
components and APIs.
• Policy makers: dashboards and district‑level insights.
Record attendance without internet; sync later.
Identifies likely dropout cases early for proactive support.
School‑owned data; aggregated analytics for districts.
Ensures efficient distribution of traffic across multiple servers for reliability and performance.
Teachers mark present/absent per class. Data can be captured on low‑end Android devices or desktops.
The risk model uses patterns (e.g., repeated absence) and prior cohort trends to estimate dropout risk.
Schools receive prioritized lists and suggested interventions; districts view aggregated trends.
OUT
AIGI
COSTECH
IDRC
PORALG
NAMA
DLI
Piloted the tool in public secondary schools; collected dropout data across ~70 districts in Tanzania mainland.
OUT hosted the study supported by COSTECH in collaboration with the President's Office – Regional Administration and Local Government.
Funded development and piloting in four private schools in Dar es Salaam.
Supported initial stages, including a district‑level model for assessing investment impact on dropout risks.
Hosted by the Open University of Tanzania (OUT). We acknowledge all partners for their foundational support.
The project is open‑source to encourage transparency, local adaptation, and sustainability. Institutions can deploy privately and contribute improvements upstream.
Use the modular packages for data collection, model scoring, and dashboards. Integrate with existing EMIS/LMS using REST or CSV pipelines.
Below is a static illustration of a risk list. Replace with your live build when ready.
| Student | Total Absences | Trend | Risk | Contributing Factors |
|---|---|---|---|---|
| Asha M. | 7 | Rising | High | Sickness |
| Peter K. | 5 | Stable | Medium | Truancy |
| Zawadi T. | 2 | Falling | Low | School-related factors |
For deployments, research collaborations, or contributions, please reach out.
Open University of Tanzania (OUT)
Dar es Salaam, Tanzania
Email 1: khamisi.kalegele@out.ac.tz
Email 2: crdo@aigi.co.tz
We gratefully acknowledge support across phases:
inschooltz