Talk by Yuriy Katser (expert in applying DS and ML in industry, co-founder of waico.tech, author of the Telegram channel @DataKatser)
The presentation consists of two parts that are closely related. The first part is devoted to the stages of preprocessing time series and problems in the data. The second part covers anomaly detection: most of the problems in the data (spikes, outliers, changes in statistical characteristics, even missing values) can be expressed as anomalies. At the end of the talk there is an overview of libraries by task.
Continue reading — it's free
Google or a one-time email code. No password.