Predictive Analytics & Scenario Simulation Engine for Education Policy in Colombia
Share project
Organization
Budget
Requested services
Overview
Empower your education investments with predictive insights and scenario simulations that drive smarter decisions, maximize impact, and transform the future of learning for millions across Colombia.
Table of contents
- The problem
- What's already in place
- Project scope
- Users and functionalities
- Success vision
- Supporting resources
The problem
Problem statement
Colombia’s education system faces persistent structural challenges, including high dropout rates in vulnerable populations, significant gaps in learning outcomes across territories, and inequitable distribution of public and private investment.
Decision-makers at all levels—national government, territorial education authorities, and private sector funders—must regularly allocate significant resources without access to forward-looking analytical tools.
The core problem is that decisions are currently based on historical data, without the ability to anticipate how the system will evolve or how different choices today could shape future outcomes.
While organizations like ExE have built robust data infrastructures, the current state of the art remains largely descriptive. Stakeholders can understand what has happened, but they cannot yet answer critical forward-looking questions such as:
What is the expected trajectory of key education indicators if current trends continue?
Which territories are most at risk of worsening outcomes over the next 3–5 years?
How would different policy or investment decisions change these trajectories?
Where should resources be allocated to achieve the greatest future impact?
As a result, resource allocation decisions remain reactive and often suboptimal, limiting the potential impact of education investments across the country.
The solution will provide the following core functionalities:
• Forecasting of Key Indicators Users will be able to explore baseline projections of education indicators (e.g., dropout rates, learning outcomes, equity gaps) over a 3–5 year horizon at national and, progressively, territorial levels.
• Scenario Simulation Users will be able to define alternative scenarios by adjusting policy, investment, or contextual inputs, and observe how these changes affect projected outcomes.
• Comparative Scenario Analysis The platform will allow users to compare multiple scenarios to support decision-making
What's already in place
Team
The project will be led by ExE’s Data and Evidence Unit, with active collaboration across technical, analytical, and strategic roles within the organization.
The core team includes:
• Project Lead (Data & Evidence Unit) Responsible for overall project coordination, alignment with strategic objectives, and engagement with Tech To The Rescue partners.
• Data and Analytics Team A team with experience in data management, statistical analysis, and education policy analysis, responsible for working closely with external experts on data preparation, model development, and validation.
• Technical Support / Data Engineering (as needed) Supporting data pipelines, integration with existing systems, and platform architecture.
• Product and Stakeholder Engagement Ensuring that the solution responds to real decision-making needs from public sector and private stakeholders, and supporting user testing and feedback loops.
ExE’s team will be actively involved throughout the entire process, not only as beneficiaries but as co-developers, ensuring strong knowledge transfer and long-term sustainability of the solution.
Project scope
To successfully develop the proposed Predictive Analytics and Scenario Simulation Engine, the following technical work and expertise are required:
- Data Preparation and Feature Engineering
Audit and preparation of existing datasets for predictive modelling
Design of data pipelines to support longitudinal and territorial analysis
Feature engineering to link education indicators with contextual and policy variables
- Forecasting Model Development
Design, development, and validation of predictive models to estimate baseline trajectories of key education indicators (e.g., dropout, learning outcomes)
Implementation of time-series and machine learning approaches suitable for medium-term projections (3–5 years)
Extension of models from national-level forecasting to territorial disaggregation
- Scenario Simulation Design and Development
Design of a parameterized framework to simulate policy and investment scenarios
Development of models that estimate how changes in inputs affect projected outcomes
Exploration of causal inference or quasi-experimental approaches where applicable
- Platform Integration and AI Interface
Integration of forecasting and simulation models into ExE’s existing data platform
Enhancement of the current AI agent to support interaction with projections and scenarios through natural language
- Model Validation and User Testing
Validation of model outputs using historical data and expert feedback
Testing with real-world policy and investment scenarios
- Capacity Building and Knowledge Transfer
Training ExE’s internal team in predictive modelling, forecasting, and scenario simulation
Documentation of models, pipelines, and methodologies to ensure long-term sustainability
Expertise Required ExE is particularly seeking support from experts in:
Data science and machine learning
Time-series forecasting and predictive modelling
Scenario modelling and simulation design
Causal inference methods (desirable)
Users and functionalities
The proposed solution is designed to serve multiple stakeholders across the education ecosystem, each with distinct decision-making needs:
-
Ministry of Education (National Level) Uses the platform to monitor the projected trajectory of key system indicators and simulate national-level policy scenarios.
-
Territorial Education Authorities (Departments and Municipalities) Use the platform to identify local risks, forecast trends in their jurisdictions, and evaluate the potential impact of different policy or program decisions.
-
Private Sector Funders Use the platform to simulate the impact of education investments (e.g., CSR initiatives) and identify where and how resources can generate the greatest measurable outcomes.
-
Researchers and Academia Use the platform to access validated models and generate evidence for policy analysis and research.
Success vision
External impact: The project will significantly improve the quality of decision-making in Colombia’s education system by enabling stakeholders to move from reactive, retrospective analysis to forward-looking, evidence-based planning.
Through forecasting and scenario simulation, decision-makers will be able to anticipate how key education indicators—such as dropout rates, learning outcomes, and equity gaps—are likely to evolve over the next 3–5 years, and understand how different policy or investment choices could alter those trajectories.
This will directly benefit:
National and territorial governments, by improving policy design and resource allocation
Private sector funders, by enabling more strategic and impactful investments
Education stakeholders, by identifying high-risk territories and prioritizing interventions more effectively
At a system level, the solution has the potential to influence decisions affecting approximately 10 million students in Colombia. Even marginal improvements in how resources are allocated—guided by predictive insights and scenario analysis—can translate into significant improvements in retention, learning outcomes, and equity.
Over time, the platform could become a core decision-support tool for education policy, helping ensure that investments are directed toward the students and territories where they can generate the greatest impact.
Internal impact: Internally, the project will represent a major step forward in ExE’s technical and analytical capabilities, transforming the organization from a provider of descriptive insights into a leader in predictive and prescriptive analytics for education policy.
The project will strengthen ExE’s internal capacity in key areas, including:
Predictive modelling and time-series forecasting
Scenario simulation and decision-support tools
Supporting resources
Fundación Empresarios por la Educación
—
Explore similar Projects
See all Projects
Sparky AI : delivering universal access to gamified & personalized education for learners worldwide
Unlock the future of education with Sparky AI, where personalized learning pathways empower every student to thrive, regardless of their background.