Data mangement EXAM 2 (ch 6&11)

structured decision

routine process made by those in operational positions

unstructured decision

no agreed upon process with a right answer, strategic level analysis by senior executives

semi structured decisions

portion with agreed upon process for what actions to take, but not all, made by middle management

intelligence (1st part of Herbert decision making process)

recognize and understand problem

design (2nd part of Herbert decision making process)

list of alternative solutions or options

choice (3rd part of Herbert decision making process)

is made and occurs when criteria is met

implementation (4th part of Herbert decision making process)

the process that gives results and feedback

speed (1st part of quality dimensions)

make a choice within a time limit

comprehensive (2nd part of quality dimensions)

narrows available opportunities according to priorties

accuracy (3rd part of quality dimensions)

decision that reflects reality of the situation

fairness (4th part of quality dimensions)

consider the interests of other parties

due process (5th part of quality dimensions)

backtracking to ensure efficiency

how to automate quality diemensions

business intelligence (data, infrastructure, analytics)

data

what you need to collect for business intelligence (NEVER DUPLICATE!!!) structure of fields, tables and primary keys

infastructure

where to put data (base or warehouse)

analytics for business intelligence

what if, goal seek, solver, scenario manager, data tables

goal seek

change singular input to see output (minimum value for average projections for 1st quarter)

Solver

changes more than one variable or restraints in formula

scenario manager

projections for estimated sales

data tables

all possibilities in one table, only interest changes in single cell

multidiemensional analysis

parameterized reports in pivot tables for certain category or regions, predictive analytics (data mining)

operational analytics

real time monitoring of business activities (internet of things, big data)

location analytics

data organized spatially (geographical information system GIS: data in the form of maps)

user interface

how user interacts with a system (data visualization)

data visualization

users see patterns from large data sets

drill downs

detailed records of aggregated totals

executive support system

senior level management for constant status of organization (sales price, customer contacts, ect.) uses dashboards or portals

key performance indicators

constant information of aggregated information (needs drilldown records)

balanced scorecard

when KPIs measure everything (needs customers and employees)

middle management decision support systems

what-if and multidiemensional analysis

AI computer system simulating human behavior

classification of larger data sets, generate solutions to complex problems, better algorthims to reach conclusions

expert system

AI recreates knowledge from base (repository where data is held) of limited info, inference engine to reach conclusion

neural network

trained to learn with layers, combined with genetic algorithm with a lot of variables, based on evolution biology

intelligent agents

preform repetitive tasks (siri, alexa)

machine learning

can preform it’s own performance (supervised or unspupervised)

computer vision

extract info from real word images (camera ceilings, shelf sensors)

database

collection of info for specific/related purpose (people, places and things)

entities

category for info (table in relational database system)

attributes

characteristics of entities organized in columns (fields of names, GPAs, majors, ect.)

record

all related attributes in the same order (rows) individual

relational databases

most common data set up in 2 dimensional tables called relations

primary keyfeild

unique identifiers for each record in table (can’t be duplicated or null/blank)

foreign keyfeild

when primary key exists in another table

relationships

organizing data between tables

data modeling

normalizing and diagram data

data normalization

eliminate redundancies and repeated groups

cardinality

a value in one entity linked to another (ex: one person can own many cars)

referential integrity

ensures data quality by preventing the deletion of records in ‘one’ table if related record exists in ‘many’ table

decision support system

software for data (defines database structure) collection and analysis

design view of database structure

able to edit fields and tables

datasheet view of database structure

where data is actually stored

property sheet

determines what kind of data can there be in each field

data manipulation

query data for insight with SQL

SQL

structured query language that communicates with databases

big data

can’t be stored/standard data management tools

volume

massive amount/quantities (90% of all data has been made in the last 2 years)

velocity

comes in every second (speed of generated data)

variety

comes from different sources/forms (videos, pictures, ect.)

database

organized for specific purpose, data never repeated, current info

data warehouse

current and historical information repository (potential interest), don’t rewrite data only update (internal and external)

data marts

data warehouse subsets (one for north america, europe, ect.)

data lake

repository for raw unstructured (mostly) data (unanalyzed)

hadoop

process for breaking down open source software framework for big data

online analytical process (OLAP)

multidimensional analytics (pivot tables)

predictive analytics

query

looks at total sales (by specific request)

data mining

cluster analysis (can’t predict without hisotical data) associations, sequences, classifications, clusters

text mining

data to predict future behavior, sentiment analysis to study inflection behavior

web mining

all mouse movements, ability to track activity better through apps