Unveiling the Determinants of Logistics Performance in the SADC Region: An Artificial Intelligence Approach

Description

This dataset supports the article "Unveiling the Determinants of Logistics Performance in the SADC Region: An Artificial Intelligence Approach". It is a balanced annual panel for 12 Southern African Development Community (SADC) member states (Angola, Botswana, Comoros, Democratic Republic of Congo, Lesotho, Madagascar, Mauritius, Namibia, South Africa, Tanzania, Zambia and Zimbabwe) over 2007-2022, giving 192 country-year observations. Mozambique, Eswatini, Malawi and Seychelles were excluded because of missing data. The dependent variable is the World Bank Logistics Performance Index (LPI, overall score, 1 = low to 5 = high). The nine explanatory variables are the Human Development Index (HDI), gross fixed capital formation (% of GDP), labour force (share of total population), industry value added (% of GDP), urban population (% of total population), exports and imports of goods and services (% of GDP), the KOF Globalisation Index, and annual growth of real GDP per capita. Data come from the World Bank World Development Indicators, the UNDP Human Development Report 2025 and the KOF Swiss Economic Institute. This is the sheet used to estimate the enhanced Radial Basis Function Artificial Neural Network (K4-RANN) and the SHAP analysis). The LPI is published only every two to three years, so most LPI values in the imputed sheet are country means. KEYWORDS Logistics performance index; SADC; panel data; artificial neural network; K4-RANN; SHAP; macroeconomic determinants; Southern Africa

Steps to reproduce

STEPS TO REPRODUCE 1. Download the LPI, gross fixed capital formation, labour force, industry value added, urban population, exports, imports, real GDP (constant 2015 US$) and population series for the 12 countries, 2007-2022, from the World Bank World Development Indicators (https://databank.worldbank.org/source/world-development-indicators). 2. Download the HDI from the UNDP Human Development Report 2025 (https://hdr.undp.org/data-center) and the KOF Globalisation Index from the KOF Swiss Economic Institute (https://kof.ethz.ch/en/forecasts-and-indicators/indicators/kof-globalisation-index.html). 3. Compute real GDP per capita (GDP / population) and its annual growth rate. 2007 is set to 0 because it is the first year of the sample. 4. Fill missing values with the mean of each country's observed values for that variable. 5. Use the "Imputed data (used)" sheet as input to the K4-RANN model, then apply SHAP to the fitted model to rank the predictors.

Institutions

• University of South Africa

  Gauteng, Pretoria

Categories

Licence