This article is a section of the article ML Feature Engineering: from Pandas to SQL (BigQuery). Time Resampling When working with time series, resampling time on a regular step is common: it allows data to be aligned for shifting rows usage, doing consistent aggregation operations, graphics, and so on. …

This article is a section of the article ML Feature Engineering: from Pandas to SQL (BigQuery). Lag Features This kind of featuring is used to encode data dynamics through time. By giving such features to a non-time series oriented model (like a Random Forest for instance), it gives the model some information…

This article is a section of the article ML Feature Engineering: from Pandas to SQL (BigQuery). Rolling windows Rolling Windows without time Feature engineering involves statistics in lot of cases, and rolling aggregations are one of them. …

Update (13/07/2022): french data gouv source has been removed from the project as the data scheme is unstable and the new format is now unusable. Seeking for data science initiatives in order to estimate french covid epidemic indicators, I came to the conclusion that there was not much on the…

Apache Airflow 2.0 release has been launched on December 17th 2020, and I wanted to try this new version to make some tests, but thinking about it, if I wanted to go to the end of the “process” and test it over the long term, I had to deploy it…

build an AI project like an IT project — A lot of articles deals with the subject of AI or data science with IT industrialization problematic. There is also a lot of tools to handle this problematic, from a mostly no-code approach (Dataiku, Rapid Miner, Knime, Salesforce Einstein Analytics, SAP Analytics, etc.) …