Optical character recognition (OCR) is a technology that transforms printed or handwritten text (or markings) into a digital dataset.

OCR can be a powerful tool when working with large amounts of physical data. It has applications in any field where physical documents play a large role, such as healthcare, banking, government administration and countless others. Recent strides in (often AI-based) OCR services have made this technique more accessible to users that may have previously lacked the resources or expertise to use it effectively.

How can OCR and OMR support decision-making in development contexts?

Human data entry is often prone to errors and can be costly and time consuming. While there are solutions for this – such as electronic data collection – this is not feasible in all contexts and applications. Some data sources must be collected manually, such as student assessments, or handwritten attendance sheets.


At Laterite, we are harnessing this technology for social impact, helping our clients to improve processes and gain valuable insights into their work.

Here are a few examples of how we are leveraging OCR (and optical mark recognition – OMR) to improve decision-making.

Studying learning loss in Rwandan secondary schools

In 2020 and 2021, we conducted paper-based numeracy assessments with over 4,000 students across 101 Rwandan secondary schools as part of the Leaders in Teaching initiative. These were conducted as part of a wider study on the impact of school closures due to COVID-19 on student retention and learning outcomes, bringing together data collected from students, teachers and school leaders.

Using OMR software, we were able to quickly digitize assessments at an accuracy rate of 99%. This allowed us to complete this process with greater speed, accuracy and cost-efficiency than if the data were to be manually entered from the paper originals, providing a high-quality data source for a study directly informing education policy in Rwanda. Read the full findings of the study drawing on this data.

Attendance monitoring for large agronomy training programs in Ethiopia

The TechnoServe East Africa Coffee Initiative collects Coffee Farm College attendance data on thousands of farmers participating in TechnoServe’s coffee agronomy training initiatives across several programs and countries. An important metric in understanding the impact of these training programs on coffee yields and livelihoods is knowing how many farmers attended training sessions. To do this, staff often monitor attendance on paper forms, which are then digitized by a data entry team. This process of manual data entry can be time consuming and prone to human error, leading to attendance reporting being often delayed.

Laterite and TechnoServe are working on a pilot program in Ethiopia where attendance data is recorded using OMR. In this process, paper answer sheets are filled out by program staff and scanned using specialized software. Through our pilot program, we hope to substantially reduce the burden on manual data entry, thus eliminating a bottleneck in the attendance monitoring process. This would lead to more accurate and timely data on program attendance, enabling better informed and faster decision-making among program managers.

Data-driven decision making for schools across Africa

Schools in low- and middle-income countries often face obstacles in obtaining information about student learning that would enable them to make decisions about how schools and classrooms are run and organized.

Laterite and Rising Academies are piloting a digitization process that uses OCR to read standardized math assessments, which are then scored and compiled into scorecards, providing fast and automated feedback to teachers and schools. To achieve this, a custom AI-based OCR model is currently being trained on student assessment data from Sierra Leone and Ghana. Once the full pipeline is operational, teachers in over 700 Rising Academies schools across East and West Africa will use smartphones to scan and upload student assessments to a cloud storage location. Scanned images are then pre-processed and fed through the OCR model, which reads student handwriting, marks the assessment, and populates school and classroom level scorecards. The scorecards are then shared with relevant stakeholders via WhatsApp. This approach aims to empower educators with real-time actionable insights to make informed decisions on student learning and curricula at the classroom, school, and district levels.

Key Takeaways

  • Optical mark recognition/optical character recognition technology has broad applications across numerous sectors and data management cases.
  • This technology can vastly improve the speed, accuracy and cost-efficiency of data entry, allowing decision making to be based on timely and high-quality data.
  • Numerous affordable and user-friendly solutions exist, allowing organizations of all sizes and backgrounds to make use of this technology effectively.

Be sure to keep an eye out for future content highlighting how Laterite is using OCR in development contexts. Get in touch with us to learn more about how this technology can support your work!

This blog was written by Oliver Budd, Research Associate at Laterite Kenya.