BigPicture BigData

The Big Picture

: basic concepts and definitions of the big data approach

Big Data _ research by private industry and government to create results and make money

Today

-   Big data is everywhere

my phone : navigation, Target app, Amazon, Facebook, real estate apps ( Zillow and Trulia ), Pandora

credit scores : determine where you can get a loan and the rate, help determine whether you can rent a 

house or apartment

insurance companies : use big data to set different car insurance rates for different individuals and cars

wearables : devices that track our sleep and exercise

-   What is data ?

Classical Psychological Approach

data is information collected from participants by asking them questions, having them complete tasks, and/or observing them

: often quantitative ( but can be qualitative )

: They usually know they are being studied.

: They usually provide informed consent ( give researchers permission to study them )

The Big Data Approach

data is any information that can be quantified

: people generally not aware that they are being studied

: tons of possibilities of what counts as data and what it can be used for

Big Data Example _ Google NGRAM

-   The Big Data Approach

Description and prediction

focus on description and prediction rather than theory

: what people do and will do rather than why they do it

Correlation

focus on measuring variables and determining the relationship between them

: rather than cause and effect

A / B Testing

experiments

: often focused on which version of a website/ad generates more clicks

: can have many versions (not just an A and a B)

N = ALL

use everyone’s data instead of having to use a sample of a population

Messy Data

messy data is okay ( because quantity is more important than quality )

: vs. little data that attempts to measure variables carefully and precisely

-   Examples of big data research

for each example, think about

1. What data was collected and how was it collected ? Did people know their data was being collected ?

2. What conclusions can (and can’t) be drawn from this research ?

3. Does using data this way improves people’s lives, make them worse, or have no effect ?

 - Walmart, Hurricanes, and Pop-Tarts

 - Farecaster

 - Obama campaign website

 - Amazon Recommendations

 - Target

 - Google Translate

 - Fitbit