Embedded Files
This section showcases the analytical dimension of my work where data, economics, climate and policy intersect. My technical experience spans the management and analysis of large, multi-dimensional datasets, combining econometric modeling, spatial analysis, and machine-learning techniques to identify relationships, assess risks, and support evidence-based decision-making. In my latest experience I focused on examining how climate and environmental change, conflict dynamics, and socio-economic systems interact to influence resilience, livelihoods, and development outcomes.
Through this work, I translate complex data into insights that guide programme design, policy planning, and strategic foresight in development and humanitarian contexts. The examples presented here highlight a range of applications, from assessing climate-security linkages and mapping vulnerability to developing analytical frameworks that inform national and regional resilience strategies.
I am now incesting my time in expanding personal technical skills in machine learning, Python programming, and advanced data science methods to enhance the analytical depth and predictive capacity of my work and to continue bridging quantitative research with actionable solutions.
Figure 1: Scatter plot showing the relationship between forced displacement and key explanatory variables, including climate anomalies, conflict exposure, and migrant characteristics. Each coefficient represents the estimated marginal effect from a linear probability model.
Figure 2: Bivariate scatter plot illustrating how the number of displaced individuals within households relates to migration intentions. The pattern highlights how household displacement conditions can influence mobility decisions.
Figure 3: Table summarizing the panel regression estimates linking climate variability (measured through drought indices) to displacement probabilities across multiple model specifications and fixed-effects structures.
Figure 4: Time-series plots showing regional variations in mobility across administrative regions over time.
Econometric and Statistical Analysis
Panel data regressions (fixed-effects, instrumental variable, Structural Equation Modeling and hierarchical models).
Micro-econometric analysis of household- and individual-level migration drivers.
Time-series analysis of conflict, climate variability, and displacement patterns.
Survey data processing, weighting, and robustness testing (Stata, R).
Data Management and Integration
Cleaning, harmonizing, and merging large multi-country datasets (Stata, R, Python).
Integration of different socio-economic, climate, and conflict data from sources such as IOM, DHS, ACLED, CHIRPS, Terraclimate, WorldPop, and FAO.
Metadata documentation, data pipeline design, and reproducible workflows
Figure 1: Illustration of the process used to match, harmonize, and merge socio-economic, conflict, and climate datasets using statistical and spatial techniques. The workflow integrates multiple data layers through coordinate matching and grid-based calculations to produce a unified analytical dataset.
Figure 1: Map illustrating the process of buffer-based conflict event counting, where the number of violent incidents is calculated within a defined radius around household survey clusters. This spatial approach helps quantify exposure to conflict and its potential socio-economic impacts.
Figure 2: Raster map displaying land use and vegetation cover across East Africa, used to validate the main livelihood and agricultural practices of surveyed populations. The visualization integrates remote-sensing data with household-level observations to contextualize exposure and adaptive capacit
Spatial and Geospatial Analysis
Mapping of conflict, migration, and environmental variables using QGIS and R.
Buffer calculation, spatial autocorrelation and clustering analysis.
Raster–vector integration and remote-sensing data preprocessing.
Visualization of multi-layered indicators through dashboards and story maps.
Machine Learning and Predictive Analytics
Feature selection and variable importance modeling (e.g., Lasso, Random Forest).
Predictive modeling of migration and displacement probabilities.
Cross-validation and model performance testing.
Figure 1: Plot showing how variable coefficients shrink as the penalty parameter (λ) increases in a LASSO regression model. This visualization illustrates the regularization process used to select the most relevant predictors while avoiding overfitting.
Figure 2: Mean-squared-error curve from k-fold cross-validation, used to identify the optimal regularization strength in the predictive model. The vertical lines represent the λ values minimizing model error and ensuring generalization.
Figure 3: Bar chart displaying the relative contribution of key predictors in determining migration and displacement probabilities. Variables with larger magnitudes indicate stronger influence within the selected predictive model.
Figure 1: Map displaying the locations of migrant departures and survey sites across East Africa. This visualization supports the spatial understanding of data collection coverage and regional migration corridors.
Figure 2: Flow diagram showing migration trajectories between countries and regions, highlighting the main corridors and relative magnitudes of movement derived from survey data.
Figure 3: Simplified diagram of a structural equation model (SEM) illustrating the interaction between climate variables, conflict exposure, and nutrition outcomes. The figure translates complex model results into an intuitive conceptual framework.
Figure 4: Circular (chord) diagram visualizing bidirectional migration flows between countries in the East and Horn of Africa. The width of each band represents flow volume, providing an immediate overview of migration intensity and interconnectivity.
Data Visualization and Communication
Production of analytical dashboards, charts, and policy visuals for reports and briefs.
Translation of quantitative findings into accessible, decision-oriented narratives.
Collaboration with design teams to ensure analytical accuracy in visual outputs.
Tools & Software
Tools & Software
Stata | R | Python | QGIS | Excel | GitHub | Power BI | ArcGIS Pro
Page updated
Google Sites
Report abuse