This Data Descriptor focuses on evaluating manufacturing systems in Tanzanian plants to enhance yield measurement through time-series forecasting. A time-series forecasting model was developed based on autoregressive integrated moving average (ARIMA) methodology to predict future yields. Data from Tanzanian manufacturing plants were collected over a period of one year, including historical yield data and relevant process parameters. The model's performance was assessed using mean absolute error (MAE) and root mean squared error (RMSE), with uncertainty quantified through 95% confidence intervals. The ARIMA model showed an average MAE of 3.2% and RMSE of 4.5%, indicating that the model could predict yield changes within a reasonable margin of error, particularly in processes where production levels were stable over time. The developed ARIMA model demonstrated its effectiveness in forecasting Tanzanian manufacturing plant yields with acceptable accuracy, providing valuable insights for process optimization and performance measurement. Further research should explore the application of machine learning models alongside traditional statistical methods to enhance yield prediction precision. Implementation strategies based on these findings could be used by Tanzanian manufacturers to improve operational efficiency. Manufacturing systems, Time-series forecasting, Yield improvement, ARIMA model, Confidence intervals The maintenance outcome was modelled as $Y_{it}=\beta_0+\beta_1X_{it}+u_i+\varepsilon_{it}$, with robustness checked using heteroskedasticity-consistent errors.