hmogasoa-a11y/SA-labour-market-2025

Analyzed Statistics South Africa QLFS microdata to identify year-on-year employment shifts across 10 industries. Identified +73k Transport and +60k Construction job gains offset by -130k Manufacturing and -122k Trade losses, resulting in net -55k jobs YoY. q4_2024_vs_q4_2025_industry_yoy.csv SA Labour Market: Q4 2024 vs Q4 2025 Industry Analysis Overview This repo calculates real year-on-year employment changes by industry in South Africa, comparing Q4 2024 to Q4 2025. The analysis accounts for Stats SA’s post-Census 2022 weight revision implemented in Q2 2025, which creates a structural break in the QLFS series. Key Finding Quarter-on-quarter comparisons within 2025 are not valid due to the new weights. This project uses Q4-to-Q4 year-on-year growth to strip out the reweighting effect and measure real employment change. Data SourcesStats SA Quarterly Labour Force Survey: Q4 2024 Microdata, Q4 2025 MicrodataVariables used: Indus, Lfs_Status/Status, WeightSource: Statistics South Africa, SuperCROSS database Methodology 1. Harmonization: Stats SA changed industry coding between quarters. Q4 2024 uses full text labels while Q4 2025 uses numeric SIC codes 1-10. Mapped using official QLFS metadata. 2. Filtering: Employed persons only. Status == "Employed" for 2024, Lfs_Status == 1 for 2025. 3. Weighting: Applied person weights Weight to get population-level estimates. NA values removed 4. Aggregation: Summed employment by 10 main industries for each quarter. 5. YoY Calculation: YoY_abs = Q4_2025 - Q4_2024. YoY_pct = YoY_abs / Q4_2024 * 100 Files FileDescription sa_labour_yoy.R Main analysis script. Commented for reproducibility / q4_2024_vs_q4_2025_industry_yoy.csv Final output table with totals and industry shares / data /Place QLFS microdata files here. Not included due to size How to Run 1. Download QLFS Q4 2024 and Q4 2025 microdata from Stats SA 2. Place files in data/ and update filenames in sa_labour_yoy.R if needed 3. Run sa_labour_yoy.R in R/RStudio 4. Output CSV exports automatically Packages dplyr, readr Key Assumptions & Caveats 1. Series break: Q2 2025 weights based on Census 2022 create non-comparability with Q1 2025 and earlier quarters. YoY Q4-on-Q4 is the earliest valid comparison post-revision. …