Combined two critical components of COVID-19 occurrence, namely temporal and spatial aspects, into a single modelling framework. Very few, if any, spatio-temporal models with reliable predictive capabilities are available in the Indian context. The proposed models are a significant addition in this regard. The framework utilises strong foundations from mathematics and statistics of epidemiological data with novel ways of analysing, fitting and predicting in real time. These models have been applied to Indian state-level data to describe and predict domestic COVID-19 situation.