Consultancy in data science, predictive modelling, remote sensing, time series and methodological support for early warning systems

Action Against Hunger in Senegal is looking to recrute the profesional services for a Consultancy for advanced analysis, development of experimental predictive models, analysis of historical time series and support for the definition of alert thresholds for the Pastoral Early Warning System. Project name: Strategic alliances for anticipatory action in response to food and pastoral crises in the Sahel. Action Against Hunger aims to hire an external consultant to strengthen the analytical capacities of PEWS through the exploration, development and documentation of advanced analysis methods, predictive or nowcasting models, historical time-series analysis and the proposal of technical thresholds, taking into account relevant temporal and seasonal dynamics, in order to support early warning, anticipatory action and humanitarian decision-making.

Formation / Diplômes

The consultancy will cover the advanced analytical components of PEWS. It will include the analysis of historical data, the exploration of predictive or nowcasting models, support for the definition of technical thresholds, analysis of the relevance of selected target variables and the production of methodological recommendations for gradual integration into PEWS information products. The consultancy will cover the advanced analytical components of PEWS. It will include the analysis of historical data, the exploration of predictive or nowcasting models, support for the definition of technical thresholds, analysis of the relevance of selected target variables and the production of methodological recommendations for gradual integration into PEWS information products. It will include, in particular:

• Review of available data sources for predictive analysis or nowcasting, primarily based on datasets and sources provided by ACF and SensingTools and complemented, where relevant, by targeted additional sources identified by the consulting firm;
• Analysis of data quality, gaps, biases, breaks in series, update frequency and spatial and temporal comparability;
• Analysis of temporal and seasonal characteristics of the available data, including differences in update frequency, seasonal cycles, lagged relationships between predictors and outcomes, cumulative effects, critical pastoral periods and the suitability of different forecast horizons for anticipatory action;
• Preparation of reproducible analytical datasets, accompanied by a data dictionary and documentation of the processing performed;
• Experimental development of statistical, machine-learning, forecasting or nowcasting models, depending on data availability and objectives validated with ACF and SensingTools, including the assessment of alternative temporal structures and forecast horizons where relevant;
• Consideration of SensingTools methodological contributions on data, models, social sensors, natural language processing and perception analysis;
• Analysis of normalized historical time series, for example through anomalies, z-scores, percentiles or other explainable methods, to support discussions on technical thresholds;
• Formulation of recommendations on thresholds, warning signals or interpretation rules that can be tested, validated or calibrated with technical teams, institutional partners or communities. This consultancy is expected to be carried out remotely.

[Click the Apply button below to apply, and Create my CV to build a CV tailored to this offer, professionally]