Business Analytics Exam 2

data visualization

process of displaying data in a meaningful way to help business managers identify insights that will support better decisions

what is raw data important for?

identifying accurate values or comparing individual numbers

what are the five storytelling guidelines for creating a storyboard?

understand the context of the data, choose appropriate visual displays, produce self explanatory visuals, eliminate clutter, use bold colors to focus the attention in certain directions.

visual analytics

combination of visualization and predictive analytics

cherry picking

using only a portion of the dataset and excluding information

requirements for a good graph

start at zero or include slashes if not, titles must be descriptive/self explanatory, use a key if necessary, no overlapping data, avoid cherry picking, do not combine graph types

when do you use slashes?

when it doesn’t start at zero and when there are interruptions in the scale

Does a histogram graph quantitative or qualitative data?

quantitative

Does a bar graph show quantitative or qualitative data?

qualitative

Does a scatterplot show 2 quantitative or qualitative variables?

quantitative

Bar graph or histogram? Large spaces between 2 bars

Bar graph

Bar graph or histogram? Thin white line surrounding bars

Histogram

smart sheet

a type of software that allows you to use dashboards

dashboard

a visual representation of a set of key business measures

BI Dashboard

special type of dashboard that is displays the status of business metrics, key performance indicators and other important data all on one screen.

What does BI stand for?

Business Intelligence

reports characteristics

static visuals or graphs, offers snapshot of historic data, can be produced using basic software (microsoft office, excel)

dashboard characteristics

interactive, includes multiple data points, includes tracking in real time, requires special software (tableau, smartsheet)

data fishing

surface data investigations to realize there is a problem that needs attention, looks for data patterns, summarizes major points from the data

drill down

looks below the surface of the data to solve a problem

data literacy

being able to understand data and communicate an accurate story about the data

ownership

one’s ability to exercise control over and limit the sharing of their own data

data ethics

the process of examining, interpreting, and applying moral principles to data

informed consent

permission given to collect or use data for stated purpose

currency

individuals should be aware when data is being sold for monetary gain by others

openness

datasets are available to the public

consent should include

purpose of the data collection, is the data confidential or anonymous, what the participant is expected to do, and the risks or discomforts involved

Is consent required when passively observing people with no interaction (Ex: people watching in a mall)

No

examples of personal data

gender, address, cell number, GPS location, internet browsing habits, tax payments

Who controls data being collected?

owner of the sensors (banks control banking data)

True or False? Consent should be given in a format that is explicit, not implied.

True

Confidentiality

The data collector knows the identity of the person but takes steps to protect that identity from being discovered by others

Anonymous

the identity of individual subjects is not known to anyone

steps to analytic process

define the problem, collect data, analyze data to find trends & patterns, build charts, graphs, or dashboards to better understand the data, ask the next question and return to collecting data.

simple random sampling

each member of the population has an equal chance of

advantages to simple random sampling

not biased. everyone has an opportunity to be in the sample (represents the whole population)

disadvantages to simple random sampling

hard to achieve with a large population

systematic random sampling

list the population, select random starting point, sample every nth person.

disadvantages to systematic random sampling

if the skip # isn’t random then the sample becomes biased

advantages to systematic random sampling

prevents clustered groups, easy to use (only have to pick random starting point and a fixed counting interval)

stratified random sampling

divide a population into nonoverlapping subgroups and then randomly select from each subgroup

advantages to stratified random sampling

produces more precise comparisons between different strata

disadvantages to stratified random sampling

May require pilot study if you don’t know strata exists, potential bias if subgroups overlap or data is outdated

multi stage cluster sampling

dividing a population into sections and randomly selecting whole sections (AKA clusters) and then follow up with stratified

Advantages to cluster sampling

More subjects can be included in a study

Disadvantages to cluster sampling

Higher sampling errors and biases

what causes sampling bias?

undercoverage, voluntary response, convenience sampling

How to prevent sampling bias?

Use random or stratified sampling to make sure that every member has an equal chance of being chosen

point of sale

data collected when a customer buys something

clickstream data

data showing what someone clicks on or views on a website

social media data

information collected from social media activity (likes, comments, shares)

sensor data

data collected automatically by a sensor/device

survey

data collected by asking people questions, usually through a questionnaire

How can a graph be misleading?

Exaggerated scales, cherry picking, pie charts don’t add up to 100%

End of an online survey, you read the following: “By submitting your answers, you agree to all terms and conditions”

consent

Google and NASA datasets are made available to anyone

openness

Stitch Fix stores customer data securely in the cloud

privacy

The https websites used by banks & credit card companies to encrypt data from customers

privacy

Customer buys a pair of shoes from shoe company A then company shares and sells data without consent with company B

currency

Patients are asked to sign a privacy form

consent

Providing software that can be used and developed further

openness

Use of third party internet “cookies” (tracking devices that remember your searches and logins)

privacy

Target uses their customer data for their own marketing campaign

ownership

A restaurant has a suggestion box for customers to leave comments as they exit

privacy

placation

calming an angry or demanding party (customer, employee) by offering refunds, adjusting terms to lower tension, etc.

Diversion

the steps companies take to hide its privacy policy or the efforts to ensure its not easy to access the privacy policy

Misnaming

when a business titles their privacy policy something misleading or names another document/article “privacy policy”.

When should you use a dashboard?

to monitor live performance and identify what is currently happening

When should you use a report?

To analyze past data and determine why did something happen