Data Analytics

business analytics

use of data by quantitative methods to give impoved insight regarding businesses

Use of analytics

Pricing - pricing goods and contracts.

Customer segmentation - advertising and promotions/data to company.

Location

Merchandise

Staffing and Personnel

Is all analytics accurate

no

Analytics strengths

Predicting broad trends

Likelihood of an event occuring

How changes to input affect likelihood

Analytics Weaknesses

Predicting individual results

Even distribution of results

Business Intelligence

The collection, management, analysis and reporting of data

Statistics

Science of uncertainty and the technology of characterisstics and patterns within data in a large database

Data mining

Focused on a better understanding of characerisitcs and patterns within data in large database

Can also find synthetic data

Data

Just numbers not valuable

ex. 77,57,95

Information

Understanding taken from data, is valuable

reliability

data is accurate and consistent

validity

data correctly measures what it claims to measure

model

reperesentation of a real system

Big data

Systems could no handle data so companies had to invest in new systems.

updated frequently

structured data

data sheet

unstructured data

photo, voicemail etc

algorithm

systematic procedure that finds a solution

Data evolution

1980 - data warehouse

2011 - data lake

2020 - lakehouse

Decision Model input

data - constant values

uncontrollable inputs - values which can change, but not be controlled

ex. weather, inflation

Decision options - valueswhich can change and be controlled

ex. pricing, staffing, investments

3 levels of modeling

descriptive modeling

predictive modeling

prescriptice modeling

Descriptive modeling

whats happening

predictive modeling

what might happen

prescriptive modeling

what is the best thing that could happen? optimization is primary goal.

optimization

process for finding a set of values which minimize or maximize a target value

prescriptive modeling constraints

limits to model that affect optimal solution

BRE Business Rule Engine

1. data collection

2. rule application

3. rule execution

4. rule based decision

Certainty of model input

stochastic or probabilitic model

deterministic model

stochastic or probabilitic model

input information is uncertain

Deterministic model/s

inputs are believe to be known

problem solvinf method

recognizing the problem

defining the problem

structuring the problem

analyzing the problem

interpreting results and deciding

implementing a solution

Uncertainty 

Imperfect knowledge of what will happen

Risk

opportunity cost

concatenation

creating strings by combiing other strings from cells or manual input

net present value

measure worth of cash flow against ‘discount rate’ over time

discount rate

opportunity cost of spending funds and/or the profit the company must recieve to justify the investment

Data set

collection of data

database

collection of related files on people, place, and things

RDB

collection of related files with keys found in other files

sorting

arranging records according to a value

filtering

displaying/ hising records based on a value

pareto analysis

AKA 80% effects are from 20% cause/s rule

useful in determining cumulative percentages

Pivot tables

custom tables built from data chosen by user

forms

allow input limits and control

Bar chart

chart which compares data, represented as bars, to show relationship between data

line chart

shows data points over time

pie chart

shows data as slices

area chart

mix of pie and line chart, showing both percentage of total data and changes over time

orbit chart

chart showing data points connected by lines in order of occurrence

bubble chart

plot 3 elements of data in 2 dimensions, bubble size represents 3rd element

combination chart

chart combining 2 sets of information, such as line and bar chart

Radar chart

chart using a spider web pattern to show multi-dimensional data elements

Stock chart

chart tracking stock prices and elements such as daily highs and lows

geographic data

map based data

color scale

cells are colored based on value

sparklines

summarize a row or column in a single cell

dashboards

visual representation fo key business values. can be ised as high level view of multiple data sources.

statistic

a summary measuring of data

metric

unit of measure

discrete metric

a metric derived from counting something

categorical data

data sorted into categories based on specified characteristics

ordinal data

data which can be ordered or ranked according to some relationship (sports ranking, class rank)

interval data

data is ranked but had constant difference between obserations and an arbitrary zero point (time, temperature)

ratio data

data which s continuous and has a natural zero point (money, student count)

qualitative data

nominals data

ordinal data

quantitative

discrete data

continuous data

frequency distribution

a table that shows the number of observations in each of several non-overlapping groups

histogram

frequency distribution in the form of a column chart

relative frequency

frequency as a portion of the total, represented as a percentage

relative frequency distribution

a tabular breakdown of relative frequencies

grouped frequency

data can be grouped into ‘bins’

group width

upper limit-lower limit/number of groups

cumulative relative frequency

the sum of relative frequencies under a particular limit

ogive

a chart for the cumulative relative frequency

kth percentile

the value at which k percent of the observations lie

quartile

quarters

quintile

5 parts

decile

10 parts

cross-tabulation/contigency table

data shows observation of data set accross 2 categorical variables.

population

all items of interest for a particular decision

sample

a subset of a population

mean

the average

median

when values are lined up, the middle value

mode

the most frequent obseration

midrange

the smallest value, added to te largest value, divided by

range

difference between largest and smallest values

interquartile range

difference between first and third quartile

variance

the average of the square of the distance of all values from the mean

standard deviation

the square root of the variance

chebyshevs theorem

for any set of data the proportions of values which lie within one standard deviation mean is at least 1-(1/k²)

process capablity index

used in manufacturing to access if parts are falling within specifications

z-score

a relative measure of the distance of the value from the mean

z-score is expressed in

standard deviation

z-score positive and negative

positive scores are to the right

negative scores are to the left

coefficient of variation

provides a relative measure of dispersion of data

coefficient of variation is useful in

determining return to risk

return to risk

the reciprocal of coefficient of variation

skewness

lack of symmetry of a data

less skew

closer to zero

positive skew

closer to one, larger right skew

negative skew

closer to negative one, larger left tail