In hydrogeological stations, missing data is a frequent issue, and filling in the missing value can be difficult. The current study presents a novel and automated method for assessing the long-term groundwater level (GwL) trend in arid and semi-arid regions using a hybrid imputation framework that effectively fills missing GwL data by preserving both spatial and temporal relationships, and comprehensive spatio-temporal trend analysis, which includes Mann-Kendall trend analysis and GIS-based spatial assessment. The entire research process was carried out in a Python environment, which approached GwL studies fully automated, scalable, more effective, and guaranteed reproducibility of the research work. The integrated approach adopted in the current study provides methodological strength and scientific novelty by combining the best-suited imputation with spatial trend analysis to better understand GwL dynamics in arid and semi-arid regions. The study used 21 years of GwL data from 42 monitoring wells for the Thoothukudi district obtained from the State Data Centre of Tamil Nadu. The temporal trends were analysed annually and seasonally using the Mann-Kendall (MK) test to detect monotonic patterns, while spatial trends were assessed using a geographical information system (GIS). The proposed hybrid imputation model effectively handled missing data while preserving spatio-temporal dependencies, achieving high accuracy with a mean RMSE value of 0.863 and a mean R-2 value of 0.793. The MK trend analysis revealed statistically significant declining trends (p < 0.05) in several wells, predominantly in the southwestern, southern, eastern, and parts of the central region of the study area. The MK trend analysis for each season identified critical stress periods during winter and summer seasons, with z-score declines exceeding -2.5 in some wells. The present study's findings emphasise the urgent need for sustainable groundwater management, particularly in the eastern, southwestern, and southern zones with pronounced declining levels.