Presentation at FlowFactor 2019. Another look at the data science hierarchy of needs specifically looking at chasing Artificial Intelligence and Machine Learning.
6. Artificial Intelligence == Machine Learning
The rapid and significant advances in AI are actually a
subset called Machine Learning (ML)
Source: https://www.guru99.com/machine-learning-tutorial.html
7. Machine Learning looks like...
● Gathering data representing
something in the world (customers,
bank transactions, t-shirts, etc)
● Defining features to describe the
data thing
● Defining the desired output
● Applying ML algorithms to find the
optimal parameters to predict similar
output given new data things
Source: https://www.guru99.com/machine-learning-tutorial.html
8. Every industry has some problem for which data science and
ML are the right tools
Retail
Personalized recommendations based on
clustering and other models to predict demand.
Just think Amazon.
Manufacturing
Machine learning is used to forecast future
demand based on inputs.
Banking
Fraud detection where you want to predict
based based on a set of features about an
individual transaction.
Even...Aviation
The efficiency airplane engines and and routes
save money and reduce environmental impact.
9. Applying ML successfully!
● ML is a tool businesses use to improve how they
operate and the services or products they provide
● This tool supports existing goals and objectives
● ML starts with same fundamental strategies
used by all data science
11. Some Fundamental Data Science Strategies
⭙ Know what problem you are trying to solve
⭙ Start simple and grow complexity over time
⭙ Practice good product development
⭙ Solve people processes along with the technology
12. ⭙ Know what problem you are trying to solve
⭙ Start simple and grow complexity over time (hierarchy of needs)
⭙ Practice good product development
⭙ Solve people processes along with the technology
Some Fundamental Data Science Strategies
14. ML lives in the predict and
optimization levels of the
hierarchy of needs
15. ML lives in the predict and
optimization levels of the
hierarchy of needs.
ML is only as good as the
data you provide it!
16. Regardless of the data
science you’re performing,
you must first gather data.
17. Then you must ensure your
data represents what you
think it represents. Data
quality is key!
18. Then you must ensure your
data represents what you
think it represents. Data
quality is key!
Again: ML is only as good
as the data you provide it!
19. To understand the data, we
must spend time inspecting
and analyzing.
20. To understand the data, we
must spend time inspecting
and analyzing.
No ML escapes defining
features for data and no
features escape this need
21. ⭙ Start simple and grow
complexity over time
Strongly consider this milestone for
your business before chasing ML
22. Businesses applying ML
Businesses spend 1-3
months to get this into
production the first time
They spend 1-3 years to
really get this right
23. Businesses applying ML
Businesses spend 1-3
months to get this into
production the first time
They spend 1-3 years to
really get this right
1-2 years to do this well
1-2 years integrate these
1+ years modeling to
integrate optimizations
24. Businesses applying ML
6+ years for ML?!
● Includes the time to integrate
data into people processes
● Tools and services exist which
do some work for you for
particular problems
● Not all businesses are starting
at the lowest need
26. Are you ready for Machine Learning?
● Do you understand the data
associations intuitively but lack
tools to make predictions?
● Do you already have quality data sources?
Are there processes in place for collection
and storage?
● Do you have people maintaining your data
ecosystem?
● Do business units already use data to
inform their decisions?
27. References and Resources
● Murat Durmus (2018) AI-Readiness – Is Your Business Ready for the use of AI?
● Guru99 - Machine Learning Tutorial for Beginners
● Rachel Schutt & Cathy O’Neil (2013) Doing Data Science: Straight Talk From the
Frontline, Sebastopol, CA: O’Reilly
● DJ Patil & Hilary Mason (2015) Data Driven. Sebastopol, CA: O’Reilly
● DJ Patil (2011) Building Data Science Teams. Sebastopol, CA: O’Reilly
● Monica Rogati (2017) The AI Hierarchy of Needs
● Nick Crocker (2014) Thirty Things I’ve Learned
● Daniel Tunkelang (2017) 10 Things Everyone Should Know About Machine Learning
● DJ Patil - Everything We Wish We'd Known About Building Data Products
28. Are you ready for Machine Learning?
● Do you understand the data
associations intuitively but lack
tools to make predictions?
● Do you have people maintaining your data
ecosystem?
● Do business units already use data to
inform their decisions?
Thank
You!
● Do you already have quality data sources?
Are there processes in place for collection
and storage?
29. Companies that merely chase the AI
breakthroughs promised in the headlines
won’t be able to deliver the real results
that will help them lead the market. Focus
first on what problem you need to solve,
and then find technology that helps.
- Gauthier Robe