Scribd
Upload
51 views15 pages

Microsoft Fabric Deep Dive - Synapse Data Science v.2

Uploaded by

Josh Townsend
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
51 views15 pages

Microsoft Fabric Deep Dive - Synapse Data Science v.2

Uploaded by

Josh Townsend
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 15

Microsoft Fabric

Guidebooks

Data Science

“Microsoft Fabric is “It is probably the


a new paradigm in biggest innovation in
how we work with data analytics since
data – it goes beyond Power BI” Mathias Halkjær
BI as we know it.” Principal Architect
Microsoft Fabric

Data Factory Synapse Data Synapse Data Synapse Data Synapse Real-Time Power BI Data Activator
Engineering Warehouse Science Analytics (coming soon)

OneLake
Collaboration
ONEL AKE
Data
Administrator
With OneLake at its core, Microsoft Fabric Data Engineer
unifies data disciplines and enhance
collaboration across all data professionals.
Data Warehouse
OneLake both ties together all the tools, Developer
experiences and technologies – and by doing
so the people working in it. OneLake

Never has it been as easy to share ones


important and impactful work instantly with
the right colleagues. Data Scientist
Data Analyst
Architecture blueprints
ALL

“One Architecture” is one of the inaugural promises of


Microsoft Fabric, that in many ways streamlines
architectural complexities.

It does, however, present a variety of


options and patterns, enabling users to
customize their experience and
maximize its potential according to
their needs.
Capacity & pricing
C A PA C I T Y DOMAIN W O R K S PA C E

Microsoft Fabric offers a variety of


purchasable capabilities divided into SKUs,
each providing unique computing power
quantified by Capacity Units (CU).

Fabric features two SKU types:

• Azure – Billed per second with no


commitment.
• Microsoft 365 – Billed monthly or yearly,
with a monthly commitment
Organization
C A PA C I T Y DOMAIN W O R K S PA C E

Warehouses, Lakehouses, Data Marts, Pipelines and Notebooks. Microsoft Fabrics launched with more new
gadgets and technologies than we could have ever dreamed of.

However, it's essential that as organizations, we don't overlook the foundational aspects such as our internal
structure, objectives, and strategic planning.

One common organizational decision ENTERPRISE HYBRID SELF-SERVICE


to consider when deploying a data
platform like Fabric, is to choose
between a centralized,
decentralized, or hybrid
implementation approach.
Microsoft Fabric

Data Factory Synapse Data Synapse Data Synapse Data Synapse Real-Time Power BI Data Activator
Engineering Warehouse Science Analytics (coming soon)

OneLake
Synapse data science

With data science in Microsoft Fabric, you can


utilize the power of machine learning features TOOLS
to seamlessly enrich data as part of your data
and analytics workflows.
NOTEBOOK
Data science is a crucial component of modern
businesses, driving informed decision-making
and predictive insights by leveraging the MODEL
power of analytics.

The Data Science experience in Microsoft EXPERIMENT


Fabric, includes Notebooks, Models and
Experiments.
DATA PIPELINES
Data science for Business Insights
Notebooks in Data Science
N OT E B O O K
“Python enables us to do
anything we want with
In Microsoft Fabric Data Science, notebooks serve as the main authoring our data – Notebooks
tool, supporting native lakehouse integration and collaboration. They can puts all that into an
be scheduled, added to workflows, and support Python and R libraries. auditable structure”

Notebooks offer integrated Spark monitoring with real-time advice from


a built-in advisor. Users can oversee Spark jobs via a monitoring hub and
leverage native VS Code integration for IDE use, working with notebooks,
Spark jobs, and lakehouses, complete with debugging support.

Jesper Frederiksen
Architect, Fellowmind
MLflow model and experiment
MODEL EXPERIMENT

Design and monitor ML models and


experiments via MLFlow.

MLFlow's robust APIs enable


straightforward handling and tracking
of models and experiments in
Microsoft Fabric.

This functionality facilitates effortless


comparison of different experiments
and automatic metric logging during
model training.
Data Wrangler
MODEL EXPERIMENT

Data Wrangler simplifies data cleaning and prep with


a user-friendly interface while maintaining Python's
robustness and reproducibility.

Its dynamic visuals, built-in analytics, and easy


Pandas data integration cater to various skill levels,
from beginners to seasoned pros.

Future features will support Spark and offer natural


language processing to code through Azure OpenAI.
Coming soon in preview

Cognitive services
MODEL

Data Science in Microsoft Fabric will incorporate Azure Cognitive Services'


ready-made AI models, including Text Analytics, Anomaly Detection, Text
Translator, and Azure Open AI's base models.

Users can leverage these models directly, with no need for Azure resource pre-
provisioning, simplifying the process of applying AI-enhanced alterations to
data in Lakehouses.
Direct impact with DirectLake
MODEL LAKEHOUSE

DirectLake, a revolutionary feature in Microsoft Fabric, serves predictions


promptly to Power BI.

By establishing a live connection to OneLake data, Power BI users can


access data in real-time. This enables data analysts to swiftly utilize the
newly published work of data scientists, augmenting real-time
collaboration and decision-making across the organization.

The unified Data Lake with OneLake and the DirectLake capability free up
the formerly isolated data workspaces, uniting data siloes and bolstering
data-driven innovation.
Get started
today

Try Fabric (microsoft.com)

Power BI Update (fellowmindcompany.com)

https://www.linkedin.com/in/mhalkjaer
https://www.linkedin.com/in/brianbonk

You might also like