Leverage an AtScale semantic layer to eliminate data wrangling, accelerate feature discovery, and publish model results to a broader audience
Data scientists rely on fast and easy access to enterprise data to be successful. The ability to access a consistent and reliable set of dimensions and metrics without manual data preparation is critical. Further, the ability to move and manipulate data with Python scripts is fundamental.
AtScale provides an analytics semantic layer solution that bridges business intelligence and data science teams. Data scientists access a curated source of data with simple Python scripts to use in ML models and AutoML platforms. AtScale delivers a clean representation of key business metrics and important analysis dimensions that that ensures consistency across all users even if underlying data sources change.
The Power of AtScale for Data Science
AtScale AI-Link brings the benefits of a semantic layer to data science users. The ability to access comprehensive data while ensuring a consistent view of critical metrics results in better models and better insights. Beyond basic metrics like revenue and shipments, AtScale supports any number of calculated metrics like average selling price or margin.
Furthermore, AtScale can manage complicated time relative metrics that are critical to time series analysis. Using simple Python scripts, data scientists can move data from AtScale into their models or AutoML platforms simplifying feature engineering and supporting consistency for production models. AtScale can keep models running by insulating them from changes to underlying data sources.
AtScale AI-Link also supports the write-back of model results through the semantic layer. This lets BI teams publish model results to analysts and managers using existing dashboard and reporting tools. Further, report consumers can leverage the AtScale dimensional model to drill down into model results as they would with analyzing historical data.
The AtScale AI-Link Advantage
- Semantic Layer Establish single view of critical business metrics (e.g., revenue, COGS, headcount) and analysis dimensions, establishing a common analytics vocabulary across all data consumers. Blend data from broader range of internal sources and 3rd party data to expand universe of features.
- Support Time Series Analysis Maintain curated set of time-relative measures with no complex SQL. Automatically create time series features based on your definitions of time.
- Feature Engineering Deliver comprehensive view of all variables with simplified transformations and minimal data engineering to feed models.
- ML Model and AutoML platform integration Leverage AtScale models with data science tools using a simple Python library and manage within your favorite notebooks.
- Programmatic Feature Creation Direct integration to consistent enterprise features and third-party data sources enable programmatic feature creation and engineering for more sophisticated models.
- Drive Visibility and Use of Predictions Automatically publish predictions within dimensional models for broader visibility and self-service consumption in existing BI tools.