Applied Econometric Data Analysis Training Course

Duration 10 days 20 modules · 65 contact hours
Level Intermediate Certificate on completion
Who attends Economists, Financial Analysts and others
Fee from USD 2,200 KSh 190,000 for Kenyan organisations
Where 52 cities + online See dates by city
Next start 9 Nov 2026 Course code SD-EE-022

Before you join: Participants should be comfortable with percentages, averages, variance, charts and basic algebra.

Dates & prices

Two start dates a month in every city, and live online on weekdays or weekends. Fees are per participant; classroom fees include tuition, materials, lunch and refreshments, and the certificate.

Suggest a date or city

Showing the next six months. Later dates, other cities and private deliveries for your team are available on request.

Course overview

Economic, financial and policy decisions often depend on data that are incomplete, unevenly measured or affected by confounding factors. Analysts need to move beyond descriptive charts and simple correlations to estimate relationships, test assumptions, quantify uncertainty and explain what results do and do not support. This course develops the practical econometric judgement needed to produce analysis that can withstand review by finance leaders, economists, auditors, regulators and programme managers.

Participants use R to prepare economic data, estimate regression models and communicate results for decision-making. The course covers data structures, exploratory analysis, ordinary least squares, statistical inference, functional forms, specification testing, heteroskedasticity, autocorrelation, multicollinearity, instrumental variables, panel-data models, time-series models, forecasting and limited dependent-variable models. Participants learn to select methods that fit the question and data, rather than treating model output as a final answer.

Delivery combines instructor briefings, guided R demonstrations, worked economic and financial cases, group analysis and structured review of model choices. Cases may use macroeconomic indicators, credit data, market prices, public expenditure, household surveys and operational finance records. Participants work in R and RStudio throughout the course and receive a certificate of completion. They leave with an annotated econometric analysis project containing reproducible code, diagnostics, model outputs, forecasts and a decision-focused findings note.

The course suits professionals who already work with economic, financial, operational or policy data and need stronger applied econometric capability. It is particularly relevant where analysis must be explained clearly to non-technical decision-makers without overstating causality or forecast certainty.

What you will be able to do

  • Prepare economic datasets in R using documented, reproducible data-management scripts
  • Estimate and interpret ordinary least squares regression models for economic questions
  • Test regression assumptions using residual diagnostics and formal specification tests
  • Select functional forms, transformations and standard errors appropriate to observed data
  • Apply instrumental-variable and panel-data methods to credible causal analysis
  • Build and assess time-series models for economic and financial forecasting
  • Estimate binary and count outcome models using logit, probit and Poisson methods
  • Produce a reproducible econometric report with findings, limitations and recommendations

Benefits of attending

For you

  • Produce credible regression analysis instead of relying on spreadsheet correlations
  • Explain coefficient estimates, confidence intervals and model limitations to decision-makers
  • Use R scripts to repeat, update and audit econometric work efficiently
  • Recognise weak causal claims before they reach management reports
  • Build a work portfolio demonstrating applied economic data analysis capability

For your organisation

  • Improve evidence supporting investment, budget, pricing and policy decisions
  • Reduce unsupported conclusions drawn from descriptive trends and correlations
  • Create repeatable analytical workflows that withstand internal review
  • Strengthen forecast development through diagnostic testing and model comparison
  • Increase internal capacity to analyse economic, finance and programme datasets

Course modules

20 modules over 10 days

Participants can organise a reproducible R project and import structured economic data.

  • R and RStudio project structure for reproducible econometric analysis
  • Script, data, output and documentation folders governed by naming conventions
  • Package installation, version checks and session information capture
  • Reading CSV, Excel and Stata files with explicit import settings
  • Inspecting object classes, dimensions, missingness and variable labels
  • Writing commented scripts that separate transformation, estimation and reporting

Participants can clean, transform and document analysis-ready economic datasets in R.

  • Data dictionaries defining units, periods, sources and variable construction
  • Identifying duplicate records, impossible values and inconsistent coding
  • Missing-data patterns assessed using summaries, plots and indicator variables
  • Recoding categorical variables with explicit reference categories
  • Creating ratios, growth rates, lags and inflation-adjusted monetary values
  • Joining datasets using keys, merge diagnostics and unmatched-record review

Participants can profile distributions and relationships before selecting an econometric model.

  • Frequency tables and summary statistics for economic variable distributions
  • Histograms, density plots and boxplots reveal skewness and outliers
  • Scatterplots with fitted lines inspect bivariate economic relationships
  • Grouping observations by geography, sector, period or demographic category
  • Correlation matrices interpreted alongside theory and data-generating processes
  • Outlier investigation distinguishes data errors from economically meaningful observations

Participants can interpret sampling variation, hypothesis tests and confidence intervals correctly.

  • Population parameters, sample statistics and random sampling assumptions
  • Sampling distributions explain uncertainty in estimated economic relationships
  • Confidence intervals reported with estimates, units and practical interpretation
  • Null hypotheses, alternative hypotheses and Type I error
  • P-values interpreted without treating arbitrary thresholds as proof
  • Statistical significance distinguished from economic and managerial materiality

Participants can estimate and interpret ordinary least squares models in R.

  • Ordinary least squares estimates conditional mean relationships between variables
  • Regression equations specify dependent variables, predictors and error terms
  • Coefficient interpretation uses units, ceteris paribus conditions and reference groups
  • R-squared and adjusted R-squared assessed without overstating explanatory power
  • Using lm() to estimate simple and multiple regression specifications
  • Presenting regression tables with coefficients, standard errors and confidence intervals

Participants can specify transformations and interaction effects that reflect economic theory.

  • Level-level models estimate absolute changes in dependent-variable units
  • Log transformations address skewness and support elasticity interpretation
  • Log-log, log-level and level-log coefficient interpretations compared
  • Polynomial terms model non-linear relationships and turning points
  • Interaction terms test conditional effects between economic variables
  • Marginal effects calculated at meaningful values of interacting predictors

Participants can diagnose common regression problems and choose proportionate remedies.

  • Residual-versus-fitted plots assess linearity and constant variance
  • Normal Q-Q plots assess residual distribution assumptions
  • Breusch-Pagan testing identifies heteroskedasticity in linear regression residuals
  • Variance inflation factors identify problematic multicollinearity among predictors
  • Cook's distance flags influential observations requiring substantive review
  • Ramsey RESET testing detects possible functional-form misspecification

Participants can report inference using standard errors suited to the data structure.

  • Heteroskedasticity-consistent standard errors protect inference under unequal variance
  • Clustered standard errors account for shared group-level disturbances
  • Comparing nested models with F-tests and adjusted fit measures
  • Information criteria support competing model comparison and parsimony decisions
  • Residual review informs re-specification without indiscriminate variable searching
  • Model assumptions documented alongside remedial estimation choices

Participants can identify threats to causal interpretation in observational economic data.

  • Causal estimands distinguished from descriptive and predictive modelling objectives
  • Directed acyclic graphs map confounders, mediators and colliders
  • Omitted-variable bias assessed through theory, timing and measurement
  • Reverse causality identified in simultaneous economic decision processes
  • Measurement error effects considered for survey and administrative variables
  • Selection bias examined when samples exclude relevant populations or periods

Participants can assess instrument validity and estimate a two-stage least squares model.

  • Instrument relevance requires a defensible relationship with endogenous regressors
  • Exclusion restrictions require instruments affect outcomes only through treatment
  • Two-stage least squares estimated using ivreg() model syntax
  • First-stage F-statistics assess weak-instrument risk before interpretation
  • Hausman tests compare ordinary least squares and instrumental-variable estimates
  • Local average treatment effects interpreted for compliers under assumptions

Participants can estimate panel models that control for unobserved unit characteristics.

  • Longitudinal data structure identifies units observed repeatedly through time
  • Balanced and unbalanced panels assessed before model estimation
  • Fixed effects remove time-invariant unobserved unit characteristics
  • Random effects assumptions compared with fixed-effects identification logic
  • Hausman specification tests inform fixed-versus-random effects decisions
  • Unit and time fixed effects estimated with clustered standard errors

Participants can design and interpret a difference-in-differences policy evaluation model.

  • Treatment and comparison groups defined from intervention eligibility rules
  • Pre-treatment and post-treatment periods establish the evaluation timeline
  • Difference-in-differences estimators compare relative changes between groups
  • Parallel-trends assumptions assessed using visual and regression evidence
  • Event-study coefficients examine effects before and after intervention
  • Policy findings reported with assumptions, uncertainty and implementation context

Participants can prepare time-series data and identify trend, seasonality and dependence.

  • Time indexes created with consistent frequency and date conventions
  • Line plots reveal trend, seasonality, structural breaks and unusual periods
  • Autocorrelation and partial autocorrelation plots identify temporal dependence
  • Stationarity assessed using visual evidence and unit-root testing
  • Differencing transforms non-stationary series into change-based series
  • Seasonal adjustment considerations for monthly and quarterly economic data

Participants can estimate, validate and communicate univariate time-series forecasts.

  • Autoregressive models use prior observations to predict current values
  • Moving-average terms represent serially correlated forecast errors
  • ARIMA model orders selected using diagnostics and information criteria
  • Training and test periods support out-of-sample forecast assessment
  • MAE, RMSE and MAPE compare forecast error performance
  • Prediction intervals communicate uncertainty around projected economic values

Participants can model dynamic relationships among economic variables and external predictors.

  • Distributed lag models estimate delayed effects of explanatory variables
  • Dynamic regression combines external predictors with autocorrelated error structures
  • Vector autoregression models interdependent economic time-series variables
  • Granger causality tests assess incremental predictive information over time
  • Impulse response functions trace shocks through vector autoregression systems
  • Structural breaks tested when policy changes alter historical relationships

Participants can estimate and interpret logit and probit models for binary outcomes.

  • Binary dependent variables require models bounded between zero and one
  • Logit and probit links map predictors to outcome probabilities
  • Maximum likelihood estimation replaces ordinary least squares fitting
  • Odds ratios interpreted carefully when communicating logistic regression results
  • Average marginal effects express predictor impacts on predicted probabilities
  • Classification tables and ROC curves assess binary model discrimination

Participants can select models for count, censored and selected economic outcomes.

  • Poisson regression models non-negative counts under equidispersion assumptions
  • Negative binomial models accommodate overdispersion in count outcomes
  • Zero-inflated models distinguish structural zeros from sampling zeros
  • Tobit models address censoring at known measurement thresholds
  • Sample-selection bias considered when outcomes appear for selected observations
  • Model choice follows outcome distribution, data process and decision purpose

Participants can review model quality and produce decision-focused econometric reporting.

  • Validation plans separate estimation, comparison and reporting decisions
  • Sensitivity analysis tests conclusions under plausible alternative specifications
  • Coefficient plots communicate estimates and confidence intervals clearly
  • Prediction plots distinguish expected values from observed outcomes
  • Regression tables include sample definitions, controls and standard-error choices
  • Findings notes state decisions supported, limitations and unanswered questions

Participants can create an auditable analysis package from source data to findings note.

  • R Markdown combines code, outputs, interpretation and technical documentation
  • Parameterised reports update recurring analysis for new reporting periods
  • Code review checklists identify hard-coded values and undocumented transformations
  • Data provenance records source files, extraction dates and transformation steps
  • Version-controlled project structure supports peer review and audit trails
  • Publication-ready charts follow clear labels, units and source notes

Participants can build, review and present a complete reproducible econometric analysis answering a defined economic or financial question.

  • Define a decision question, outcome measure and analytical population
  • Prepare source data using documented cleaning and transformation scripts
  • Select an econometric model justified by theory and data structure
  • Estimate competing specifications and review diagnostic evidence
  • Present findings, uncertainty, limitations and decision implications
  • Defend modelling choices during peer review and stakeholder questioning

Exercises and examples are adapted to each cohort's sectors and roles.

What you take back to work

  1. Reproducible R econometrics project template
  2. Economic data quality checklist
  3. Regression diagnostics decision tree
  4. Causal inference model-selection guide
  5. Time-series forecast validation template
  6. Annotated econometric findings report

Who should attend

  • Economists — need defensible estimates and forecasts from observed economic data
  • Financial Analysts — need to quantify drivers of financial performance and risk
  • Policy Analysts — need to assess programme effects and policy relationships
  • Risk Analysts — need models for credit, market and operational risk indicators
  • Monitoring and Evaluation Specialists — need stronger evidence from survey and programme data
  • Accountants and Finance Managers — need to interpret econometric evidence used in planning decisions

Skills and tools covered

R data preparationRegression diagnosticsCausal inferencePanel-data modellingTime-series forecastingEconometric reporting

Tools and standards

R, RStudio, tidyverse, fixest

How the course runs

Training methodology

The course uses short instructor briefings followed by guided work in R and RStudio. Participants analyse realistic economic, financial and policy datasets, progressing from data preparation to estimation, diagnostics and reporting. Group exercises focus on model selection, interpretation and challenge of assumptions. Simulations show how omitted variables, measurement error and serial correlation affect results. Each participant develops an application plan for a current workplace question and receives structured feedback on their final econometric analysis.

Requirements and prerequisites

Participants should be comfortable with percentages, averages, variance, charts and basic algebra. Prior exposure to regression, statistics or R is helpful but not essential; the course introduces required R workflows and refreshes core statistical concepts before model building. Participants should have experience working with spreadsheets, databases or structured business data and should be prepared to interpret quantitative outputs. Advanced calculus, prior programming experience, formal econometrics study and previous use of specialist statistical software are not required.

A typical training day

08:30 – 10:30First session
10:30 – 10:45Refreshment break
10:45 – 12:30Second session
12:30 – 13:30Lunch and networking
13:30 – 15:00Third session
15:00 – 15:15Refreshment break
15:15 – 16:30Workshop and daily review

Live online deliveries follow the same structure in East Africa Time, with shorter screen blocks and longer breaks.

What the fee includes

  • Instruction by a practitioner facilitator
  • Full course workbook and materials
  • Exercise files, templates and case studies
  • Refreshments and lunch on classroom deliveries
  • Certificate of completion
  • Post-course application plan

Frequently asked questions

Participants need comfort with basic numerical information, charts and structured datasets. Prior R or econometrics study is useful but not required; the course builds required foundations before progressing to advanced models.

Participants should bring a laptop capable of running R and RStudio, plus a current workplace question if they have one. Training datasets are provided, and participants may use non-confidential work data where employer policy permits.

It is designed for economists, finance analysts, policy analysts, risk specialists and monitoring and evaluation professionals. It also suits managers who regularly commission or review quantitative analysis and need to challenge model assumptions.

This course focuses on estimating economic relationships, assessing causal claims and producing forecasts using econometric methods in R. It spends less time on general dashboard design or introductory statistical theory, and more time on model specification, diagnostics and interpretation.

Participants can use the project structure, diagnostic checklist and reporting templates for forecasting, policy evaluation, credit analysis, pricing analysis or expenditure studies. The capstone format provides a direct model for a workplace analysis project.

The first part establishes R workflows, data preparation, regression and diagnostic judgement. The second part develops causal, panel, time-series and limited-outcome models, ending with a capstone analysis review and presentation.

Questions about this course?

Ask about dates, fees, group bookings or running it for your team. Our training team replies by email.

Ask on WhatsApp info@skillsetdevelopment.com

Related courses in Economics & Econometrics