Time series forecasting with Imputation for Methane Emission Monthly data

This study performed a comparative approach to forecast methane emissions from the annual data from 2019-2021 which was gathered from sentinel-5P satellite images and used famous time-series model PROPHET model. We used NAÏVE forecasting model as baseline model to compare performance of the Prophet model. The models used that were evaluated using performance metrics that gave MAE 20.5. Using this best model, average methane emissions were also forecasted from 2022 to 2024. The idea of this post is to use a univariate time-series dataset and produce a best-fit model that allows us to confidently predict future production.