Data Science in Microsoft Fabric

Products

Dynamics 365 ERP

Dynamics 365 Finance
Connect your operations with your finance

Dynamics 365 Supply Chain Management
Take control of your supply chain

Dynamics 365 Project Operations
Unify all your project management

Dynamics 365 Human Resources
Optimize your HR processes

Dynamics 365 Commerce
End-to-end retail management solution

Dynamics 365 Business Central
Unified, adaptive business operations platform

Dynamics 365 CRM

Dynamics 365 Sales
Turn your opportunities into sales

Dynamics 365 Customer Service
Build customer loyalty

Dynamics 365 Customer Insights
Connect with your audience

Dynamics 365 Customer Voice
Know your customers

Dynamics 365 Field Service
Optimize your field service

Power Platform

Power Pages
Connect with your customers and suppliers

Power BI
Base your decisions on data

Power Automate
Automate your processes

Power Apps
Turn your ideas into applications

Power Virtual Agents
AI-driven, conversational support bots

Modern Work

Microsoft 365
Your office in the cloud

Microsoft Fabric

Data Factory
Automated, flexible, real-time data orchestration

Data Engineering
Efficient and scalable data processing

Data Science
Unified, automated, end-to-end modeling

Data Warehousing
Scalable, integrated, real-time warehousing

Real-Time Intelligence
Intelligent data insights in real time

OneLake
Unified data lake platform

Data Activator
Real-time automated data triggers
Blog
Contact us

Data Science in Microsoft Fabric

Empowering Data Science with Unified Insights

Powering the Full Data Science Journey in One Platform.

Data wrangler for interactive data preparation

Data Wrangler provides an intuitive interface for data scientists to explore, clean, and transform datasets. It offers dynamic data displays, built-in statistics, and visualization tools, enabling users to perform data wrangling tasks with minimal coding. Additionally, Data Wrangler generates corresponding Python code for each operation, promoting reproducibility and collaboration.

Seamless integration with notebooks

Integrated platforms often offer tools for creating visual representations of data, such as histograms, scatter plots, and box plots. These visualizations help in identifying patterns, trends, and anomalies within the data, facilitating a deeper understanding and aiding in the formulation of hypotheses for further analysis.

Built-in data profiling & quality assessment tools

This feature includes tools that automate the detection and correction of data quality issues, such as handling missing values, correcting inconsistencies, and standardizing formats. Automation streamlines the data preparation process, reducing manual effort and minimizing errors, thereby accelerating the overall data analysis workflow.

Integrated MLflow for experiment management

Fabric incorporates MLflow, an open-source platform for managing the machine learning lifecycle. This integration enables data scientists to log parameters, code versions, metrics, and output files during model training, ensuring comprehensive tracking and reproducibility of experiments.

Collaborative notebooks with real-time editing

Fabric provides collaborative notebooks that support real-time co-authoring, allowing multiple users to work simultaneously on data exploration, analysis, and modeling tasks. This feature enhances team productivity and fosters a collaborative environment for data science projects.

Centralized model registry & version control

Fabric offers a centralized model registry that facilitates the storage, versioning, and management of machine learning models. This centralized approach ensures that team members can access the latest model versions, track changes, and maintain consistency across deployments.

Integration with SynapseML for scalable machine learning

Fabric incorporates SynapseML, an open-source library that simplifies the creation of scalable machine learning pipelines. Built on Apache Spark, SynapseML enables distributed training and evaluation of models across large datasets, facilitating tasks such as anomaly detection, computer vision, and text analytics.

Predict function for batch scoring

Fabric provides the PREDICT function, which allows users to operationalize machine learning models by performing batch scoring across various compute engines. This function supports models saved in the MLflow format and enables the generation of batch predictions directly from notebooks or model item pages, streamlining the deployment process.

Seamless integration with Power BI for real-time insights

Fabric’s integration with Power BI enables the visualization of prediction results in real-time. By leveraging Direct Lake mode, users can access the latest predictions without the need for data loading or refresh, facilitating timely and informed decision-making.