Data Analytics Midterm

Difference Between Human Brain and Computer

human brain changes physically as you live and learn

computer is structured physically the same as it is programmed, it only works in the way it is programmed to work

Volatile Memory

power goes out, whatever is in there is gone, really fast, billions of a second

Non Volatile Memory

storage, keeps memory even when the power goes out, millions and thousands of a second

Binary Code

all numbers and text/characters are expressions of 0s and 1s

Types of Business Analytics

descriptive, predictive, prescriptive

Descriptive Analytics Questions

what happened? what is happening?

Descriptive Analytics Enablers

business reporting, dashboards, scorecards, data warehousing

Descriptive Analytics Outcomes

well defined business problems and opportunities

Predictive Analytics Questions

what will happen? why will it happen?

Predictive Analytics Enablers

data mining, text mining, web/media mining, forecasting

Predictive Analytics Outcomes

accurate projections of future events and outcomes

Prescriptive Analytics Questions

what should I do? why should I do it? how can we make it happen?

Prescriptive Analytics Enablers

optimization, simulation, decision making, expert systems

Prescriptive Analytics Outcomes

best possible business decisions and actions

Diagnostic Analytics

why did it happen?

Key Performance Indicators

KPI

ex. how much you’ve made, employee turnover rate

Data Sources

these extract data

Enterprise Resource Planning

ERP

is a data source

Point of Sale

POS

data source, information of items, price, customers, etc. can interact with accounting and financing systems for business operations

Online Transaction Processing

OLTP

data source, processes your transactions for you

Extract Transform Load

ETL

extract data from sources, transform data, and load it into the enterprise data warehouse

Metadata

data about the data (ex. what do the columns mean, where does the data come from)

formats include XML, JSON, YAML, CSV

Replication

back your stuff up

Data Analysis Process

ask, prepare, process, analyze, share, act

Act

step 1 of the data analysis process

Step 1: Ask

ask effective questions, define the problem, use structured thinking, communicate with others

Prepare

step 2 of the data analysis process

Step 2: Prepare

understand how data is generated and collected

identify and use different data formats, types, and structures

make sure data is unbiased and credible

organize and protect data

Process

step 3 of the data analysis process

Step 3: Process

create and transform data, maintain data integrity, test data, clean data, verify and report on cleaning results

Analyze

step 4 of the data analysis process

Step 4: Analyze

use tools to format and transform data, sort and filter data, identify patters and draw conclusions, make predictions and recommendations, make data driven decisions

Share

step 5 of the data analytics process

Step 5: Share

understand visualization, create effective visuals, bring data to life, use data storytelling, communicate to help others understand results

Act

step 6 of the data analysis proces

Step 6: Act

apply your insights, solve problems, make decisions, create something new

Bias

distortion of data has a standpoint (ex. search up lobster on google, most images are it as food and not a living creature)

Data

“things given”, single recorded fact

Modern Business Data Definition

digitized records or evidence, facts about the world that have been captured in a form a computer can store and process

Taxonomy

a classification system that sorts data by how it is organized and what it measures

has two layers → how the data is structured and what kind of values structured data contains

Structured Data

organized into tables with rows and columns, spreadsheet format or relational database

each row is a record and each column is an attribute

values are either numeric or categorical

Semi-Structured Data

does not fit nearly into rows and columns but is not a free for all either

carries tags and markers that label its contents

ex. XML, JSON, HTML, log files

Unstructured Data

has no predefined organization at all

ex. text documents, images audio, and video

the largest share of what organizations collect and is historically the hardest to analyze

Categorical Data

places each record into a group or class, includes nominal and ordinal

Numerical Data

measures quantity, includes interval and ratio

Nominal Data

categorical, labels with no inherent order (ex. ZIP code, department)

Ordinal Data

categorical, has meaningful order (ex satisfaction scale, freshman/sophomore/junior/senior)

Interval Data

numerical, can have negative values, no decimals (ex. temperature)

Ratio Data

numerical, zero is absolute/no negative values, can have decimals (ex. age, weight)

Why Taxonomy Matters

storage and tools, valid analysis, business value

Six Types of Problems

1. making predictions

2. categorizing things

3. spotting something unusual

4. identifying themes

5. discovering connections

6. finding patterns

Comma Separated Value

CSV

text file, unstructured data

80%

how much data is unstructured

Specific

s of smart

Measurable

m of smart

Action Oriented

a of smart

Relevant

r of smart

Time Bound

t of smart

Boolean Data

true or false, 0 or 1

Geolocation

categorical as a state or country, numerical as longitude and latitude

String

text

Graphical User Interface

GUI

Continuous Data

process of measurement, can have decimals (ex. weight)

Discrete Data

counting things, whole numbers (ex. number of items)

Data Transformation

changing around a data set, prep and process

Pivot

takes rows and puts them into columns, structures your data differently

Personally Identifiable Information

PII

Types of Dirty Data

duplicate, outdated, incomplete, incorrect/inaccurate, inconsistent

Duplicate Data

any data record that shows up more than once

Outdated Data

any data that is old should be replaced with newer and more accurate information

Incomplete Data

any data that is missing important fields

Incorrect/Inaccurate Data

any data that is complete but inaccurate

Inconsistent Data

any data that uses different formats to represent the same thing