Yesterday, a colleague sent me a paper. Topic: Gap filling methods. Application: Daily precipitation time series in Oman. A fusion model promised an extreme quality boost compared to any single model. Since gap filling and time series extension of meteorological data is more or less the core of my job, I was genuinely exited. The paper promised an increase in R² from 0.72 (gradient boosting, best single model) to 0.88 (fusion model). #RStats #python #meteo #gapfilling #ML #AI #science #paper 1/X