Datrics updates

Product updates v1.2.11:
stability, performance and improved binning.

Binning improvements

We significantly improved the Binning functionality of UX:
  1. Custom configuration. The user may save the custom configuration and apply the rules at the pipeline full-run without the necessity to apply changes for each column separately.
  2. Bins bound conditions. The user has a possibility to specify the bins' bound conditions using the strictness of the boundaries (like "included" or "not included" )
  3. Outliers treatment. Now the outliers treatment strategy is manually configured - the user has a possibility to specify the behavior regarding the outliers - keep or remove them.
Freezing in Parse List and Parse JSON

The functionality for working with JSONs and Lists now has the freezing option. This allows to train the pipeline to produce set number and names of features, even with the new data that's different to the previous data set. Working with deployed pipelines that consist of those bricks had become easier.
Text description of the left sidebar

The left sidebar underwent the complete rework. Now it's much easier to navigate and find the bricks you need. You can also find a detailed description of each brick.
BigQuery connector

You can pull data from google big-query in Datrics. Just upload your credentials.json file.
Possibility to copy pipeline into the new pipeline group

Now you can copy your pipeline not only into the pipeline group but also in the new project or even in the new group, which makes sharing and reuse of pipelines easier.
Improvement of date-time type conversion

We have extended the functionality of Typization and Parse Datetime bricks by adding the possibility to directly choose the format for the input date-time columns.
Read more about our other updates and features and use-cases!

The demo is accessible here.
Do you want to discover more about Datrics?
Drop us a note to follow-up on any issues, support requests or enterprise plans.
Your feedback matters!