The datasets are used to test an embedding imputation model. There are two different experiments: finance and mobile applications.