Data Science 101

DATA SCIENCE?
what is
DATA SCIENCE?
- making an impact with data -
Ta Virot Chiraphadhanakul
Data Scientist, Facebook

Harvard Business Review (October 2012)

VALUABLE INSIGHTS
transforming data into

“No one should die because they cannot afford
health care and no one should go broke because
they get sick. If you agree please post this as your
status for the rest of the day”
Information Evolution in Social Networks (Adamic et al. 2014)

Social Influence in Social Advertising: Evidence from Field Experiments (Bakshy et al. 2012)

A 61-million-person experiment in social influence and political mobilization (Bond et al. 2012)

DATA PRODUCTS
transforming data into

email classification
spam detection

INTERESTING STORIES
and sometimes, transforming data into

http://www.google.org/flutrends/

Coordinated Migration (Facebook Data Science page)

https://jawbone.com/blog/2014-year-review/

How families interact on Facebook (Facebook Data Science page)

making data accessible to everyone
OPEN DATA

http://fragile-success.rpa.org/maps/jobs.html

don’t have the data you want? just ask for them!
CROWDSOURCING DATA

hack-a-thons, meet-ups, competitions, fellowships
BUILDING COMMUNITIES

https://seanjtaylor.github.io/out-for-justice/

http://ivory-infinity-763.appspot.com/

hubway.virot.me bayareabikeshare.virot.me

http://drewconway.com/zia/2013/3/26/the-data-science-venn-diagram

data wrangling (aka data janitor work)
BIG DATA, BIG PILE OF JUNK(?)

Harvard Business Review (April 2014)

ice cream consumption vs. murders
CORRELATION ⇏ CAUSATION

randomized controlled experiments, a/b testing
CAUSAL INFERENCE

regression, classification, clustering, collaborative
filtering, etc.
MACHINE LEARNING TASKS

feature engineering
GARBAGE IN, GARBAGE OUT

cross-validation and
penalizing model complexity (regularization)
AVOID OVERFITTING

measure the right things
METRIC, METRIC, METRIC

Seven Pitfalls to Avoid when Running Controlled Experiments on the Web (Crook et al. 2009)

speed, simplicity, cost of obtaining data, …
BEYOND MODEL ACCURACY

Graphs in Statistical Analysis (Anscombe 1973)

EXPLORATORY DATA ANALYSIS
summarize and visualize important characteristics of a
data set

http://bl.ocks.org/mbostock/4063663

THANK YOU
Ta Virot Chiraphadhanakul
http://ta.virot.me | @tvirot
THANK YOU

Data Science 101

More Related Content

What's hot

Similar to Data Science 101

Recently uploaded

Data Science 101