Yo. Big Data understanding data science in the era of big data. Natalino Busa @natalinobusa
May 12, 2015
Yo. Big Dataunderstanding data science in the era of big data.
Natalino Busa@natalinobusa
Parallelism Mathematics Programming
Languages Machine Learning Statistics
Big Data Algorithms Cloud Computing
Natalino Busa@natalinobusa
www.natalinobusa.com
Understanding Big Data
What is life?
Why are we?
What is reality ?
● (almost) everything is a number
● A few guys came with some good ideas: Aristoteles, Galileo, Popper, Fisher, Pearson, Bayes
What has changed in 2500 years?
Aristoteles
Analytical reasoning
induction
deduction
Causality
Ontology
Galileo
Scientific method
experiment
reproducibility
math formula’s as models
Popper
Falsification
Exact sciences
Models have to adhere reality
Statistical inference:
Can we falsify beliefs?
Pearson
Statistical method
Null hypothesis
hypothesis testing
Principal Component Analysis
Correlation Coefficient
Fisher
Statistical method
Likelihood function
Significance
Distribution
Sufficient statistics
Bayes
Math of belief
belief inference
network of beliefs
hypothesis -> beliefs
What about it?
The shocking truth:
1) we use these concepts every day
2) we have a pre-scientific intuition of these ideas
Why do we bother?New problems are related to understanding human behavior:
understand needs, desires, dreams, ambitions, cravings, and hopes.
Models have a great side effect: they help us predicting the future.
three weapons:Processing power: Models becomes faster: can unroll for everybody’s profilesSources: extract more data features, use different data.Context: exploring information in order to understand the person.
So, why data?
Data is our way of understanding life and reality.
How to deal with it?
Well, it’s quite simple, in a nutshell:
This is what (data) science is about:
data -> hypothesis -> validation
… but what we (mostly) really do is:
Use very little data
-> apply it to pre-formulated beliefs
-> come up with some “gut feeling”
Validate it:
It didn’t work? “Well, I am still right. ”
Just buy the damn’d thing.
What’s the problem with it?
● Context○ we could use some more data○ insufficient feature engineering
● Add more hypotheses○ we could explore more scenarios, “pivoting”○ look at the problem from other angles○ need data “artistry”
Big data to the rescue?
Big Data is the domain which:
transforms numbers to insights
services to experiences
Big data to the rescue?
by aggregating data sources across users across applications across domains
Big data to the rescue?
in order to providing personalized and relevant results
to the consumer of the given service anywhere, anytime.
Some small headaches
users != consumers
N=all : doesn’t mean you don’t need to clean it
Not all data is born equal
you don’t know what you don’t know
Keep exploring.
Your problem might not be captured by your data features.
Some small headachesTough to inspect big data.
Tough to reason about big data.
representativity/bias, support, and segmentation
signal to noise ratio:
look at GFT (Google Flu Trends) for instance
Diminishingreturns
Most of models pretty good after a few weeks
winner added just about 5% moreafter 1 year, 300 ensemble model
moral:move on, get a new angle
How to compare?You know the answer (supervised methods)
confusion matrix
ROC (Receiver Operating Characteristic)
Mean Square Error (MSE)
You don’t know the answer (unsupervised methods)
objective function
access ground truth
A/B testing
Which is right?
Beware the modeling risksOverfitting train data
Not enough “support” in the population
Not enough features available/discovered
Not well defined objective function
Object functions
“ you can please some of the people some of the time”
Object functionsMany want a slice of the cake when it’s about object functions
● what the user wants
● what the community wants
● what marketing wants
● what business wants
● what finance/monetization wants
Data scientistsData artists,Data analystsData scientistsData engineers
confirmatory analysis: domain knowledge, statisticians and data analysis
exploratory analysis : data artists/scientists
operational analysis: data engineers , data technologists
When is data science cool?
What do we look in the haystack?outliers
outliers are indicators and/or noise
groups
(Similarity metrics, PCA, SVD)
Big data as pragmatic approach to:
cheap storage
distributed computing
How to enjoy and compare data science?
enjoy the artistryappreciate the genius
cross-validationavoid falling into the trap of over-fitted models
define baselineavoid qualitative methods
define a metric, put the models to the bench, compare results
Parallelism Mathematics Programming
Languages Machine Learning Statistics
Big Data Algorithms Cloud Computing
Natalino Busa@natalinobusa
www.natalinobusa.com
Thanks !Any questions?