Upload
others
View
9
Download
0
Embed Size (px)
Citation preview
Microsoft BI Consultant
Ordina Belgium
Mulkens Jan Hermans Kimberly
Data Scientist
Ordina Belgium
Democratizing Data ScienceWith Microsoft
Who am I
Kimberly Hermans
Data Scientist & CRM consultant
Practice Manager: Data Talks
“Bringing big data, data science & visualisation together and
getting rid of the silos is what will get you the return on data”
Who am I
Jan Mulkens
Microsoft BI Consultant
Competence Lead: Microsoft Advanced Analytics
@JanMulkens
www.janmulkens.be
www.globalpowerbi.com
ActivitiesBlog:
- www.janmulkens.be
- www.globalpowerbi.com
Speaker:
- Belgian SQL Server user group (DataMinds)
- Belgian Information Worker user group (BIWUG)
- Various other external events
- Webinars
- Guest lectures
Organiser:
- Internal Ordina events
- Virtual user group events
Our employer
What Ordina says
“We increase our customers 'Return on Data' by taking them on a journey to a modern &
innovative data culture. We organically grow into the most focused, fast, flexible, friendly and fun BI & Data Science community on the Belgian
Market employing >100 growing & happy employees”
What we experience
“We increase our customers 'Return on Data' by taking them on a journey to a modern &
innovative data culture. We organically grow into the most focused, fast, flexible, friendly and fun BI & Data Science community on the Belgian
Market employing >100 growing & happy employees”
*Emphasis is my own
Takeaway
Setting expectations
• Targeted at anyone looking to promote data science within their organisation. Including data scientists looking to have the business build upon their work
• Focus is on learning about which tools can help enable citizen data scientists
Tools enable citizen data science.
Microsoft made the tools for you.
What to expect Take-away
End User Everyone
The age of
Classic BI
End UserIT Analyst
The age of
Self Service BIThe age of
Data Culture
Data Culture
Amir Netz & Kamal Hathi on the Age of Data Culture
https://www.youtube.com/watch?v=rg3lzHRxNqM
Agenda
Intro to Data Science A Fool with a ToolTooling
Intro to Data Science
Intro to Data Science
Why?
Value
Source: Gartner (October 2014)http://www.gartner.com/newsroom/id/2881218
Why?
Source: Gartner (October 2014)http://www.gartner.com/newsroom/id/2881218
Why?
Value
Source: Gartner (October 2014)http://www.gartner.com/newsroom/id/2881218
Why?
Value
Who?
ML
Researcher
Data
ScientistML Engineer
Citizen Data Scientist
Up to15hrs / week
Up to15hrs / week
Up to15hrs / week
Full work week
Up to 4 hrs / week
Source: Generalized from O’Reilly’s “2015 Data Science Salary Survey” (sep 2015)
ETL Data Cleaning Machine LearningExploratory Data Analysis
Data Science Time Schedule
Citizen Data Scientists- Exploratory Analysis
- Visualization
- Putting insights into action
Come from both the business and IT
- BI Developers & Analysts
- Power users
Why?- Requirements going beyond BI
- Shortage of data scientists
- Ad-hoc analysis doesn’t scale
- Getting direct support from the business and IT for your data science project
- Bigger chance at project success & deployment because of buy-in
How?Education
- Basic education on required data science topics where necessary
EducationDifferent data problems require different knowledge
Clustering & dimensionality reduction: (e.g. K-means clustering)
Regression: (e.g. Linear regression)
Association: (e.g. Recommenders) Classification: (e.g. Logistic regr., trees)
Machine learning algorithms
Likely buy
BuySimilar
EducationDifferent data problems require different knowledge
Predict
values
Find
unusual
occurrences
Discover
structure
Predict
categories
EducationPick a data science process as a guide
Define
Goal
Collect
Data &
explore
Build
Model
Evaluate
Model
Present
Results
Deploy
Model &
monitorKDD
CRISP-DM
CCC Big Data Pipeline
Microsoft Team Data Science Process
...
How?Education
- Basic education on required data science topics where necessary
Create a Data Culture
- Open communication
- Supportive learning environment
- Reviews of models & performance
- Data scientist available to help
How?Education
- Basic education on required data science topics where necessary
Create a Data Culture
- Open communication
- Safe learning environment
- Reviews of models & performance
- Data scientist available to help
Tooling
- Using the right tools to help non-data scientists develop predictive solutions
Tooling
Tooling
Data Catalog- Register
- Annotate
- Understand
- Discover
Ingest Prepare Analyze Publish Consume
Tooling
Cognitive Services
Tooling
Azure ML Studio- Easy to get started
- Create, Share & Publish
- Supports team collaboration
- Web-Basedhttps://studio.azureml.net
Cortana Inteligence Gallery
Cortana Inteligence Gallery
Cortana Inteligence Gallery
Cortana Inteligence Gallery
Cortana Inteligence Gallery
Tooling
Power BI
Azure DS VM- Microsoft R Server
- Anaconda Python
- Julia Pro
- Jupyter Notebooks (R, Python, Julia)
- Visual Studio Community Edition- With Python, R, node.js tools
- Power BI Desktop
- SQL Server 2016 Developer Edition- Including support for in-database analytics with R Server
- Open Source Deep Learning Tools
What could possibly go wrong?
A Fool with a Toolss
Horror stories (logic)
Horror stories (stability)
Horror stories (relativity)
A B
Source: http://timoelliott.com/blog/wp-content/uploads/2015/03/citizen-data-scientists.png
Wrapup: Be a citizen data scientist
Wrapup: Use tools that enable you
Thank you!
Resources
ResourcesGet inspired: Amir Netz & Kamal Hathi on “The age of data culture”
- https://www.youtube.com/watch?v=rg3lzHRxNqM
Gartner on Advanced Analytics (Oct 2014)- http://www.gartner.com/newsroom/id/2881218
Azure Data Catalog- https://azure.microsoft.com/en-us/services/data-catalog/
Cortana Intelligence Gallery- https://gallery.cortanaintelligence.com/
ResourcesPower BI Partner Showcase
- https://powerbi.microsoft.com/en-us/partner-showcase/
Azure Data Science VM- https://azuremarketplace.microsoft.com/en-
us/marketplace/apps/microsoft-ads.standard-data-science-vm
Machine Learning Basics Infographic- https://docs.microsoft.com/en-us/azure/machine-learning/machine-
learning-basics-infographic-with-algorithm-examples
Microsoft Professional Program in Data Science- https://academy.microsoft.com