Data Analytics Exam 2 Review

ETL

Extract, Transform, Load

Gestalt Principles Examples

Proximity Similarity Enclosure Connection Symmetry Continuity Closure Figure and Ground

Data-to-ink ratio

Tufte principle, ratio of data to ink should be maximized non-data ink and redundant data ink (chart junk) should be removed essentially less is more remove backgrounds, redundant labels, borders, and bold effect, lighten labels, lighten or remove lines, label directly

warehouse v. production

Warehouse - storage for older files Production - day to day files Configure your data by having Power BI workbooks work off a warehouse rather than production

human element

critical thinking and judgement

(T/F)

dashboards are constructed from visualizations

bottom-up

data drives your research, you know the scope but you aren't sure what's in the data

Card

shows any number, such as a sum or an average

text tables

simple text table that shows data along two columns in tabular format Don't have too many columns in a text table

canvas

space where you design your dashboards

pre-attentive processing

tuned to detect a special set of visual attributes, which results in certain elements standing out - all without conscious thought

measures

typically numeric and are considered a dependent variable, value is a function of one or more dimensions e.g. units, unit price, profit

Tables

a grid with rows and columns, useful for comparing values of one or more categories

'NUMBER OF ITEM TYPES'

a measure that determines how many different products there are

development of dashboards

a process that relies on different types of knowledge

instance

a specific object e.g. Vincent Van Gogh is an instance of Painter Honda Pilot is an instance of Car

drillthrough

connect two or more reports that have the same content

Tufte Principles: Departure

"Simple is Better" depends on your audience An audience with a significant stake in the subject matter often demands precision with lots of details

color

- no color overload - be consistent and effective when using color - cater the color for your audience - understand color context

Analytics Lifecycle

1. Define scope and approach 2. Data Identification & Gathering 3. Data integration and Mapping 4. execution of Analytics 5. Visualization and Reporting 6. Present Findings & Enhance

Analytics Life Cycle

1. Request data 2. Receive data 3. Perform pre-load checks 4. Load data into data analysis platform 5. Perform post-load checks 6. Add custom fields 7. Perform Analytics 8. Export results/data

Elements of Visualization

Data colors Size Data Label Plot Area Tooltip Buttons Text Box Image Shapes

"manage relationships"

Here can add or delete a link

Storytelling approach

Memorable Relatable Lead to action

(T/F) A good dashboard should make your point without the need for words or explanations

True

Edit Queries

allows the analyst to manipulate the data using a powerful calculation engine with ability to do arithmetic, string, date, aggregation, and other operations support for Boolean logic and conditional statements Add a new column to date to enhance visualizations

Get Data

button used to get to data connectors such as text/CSV, Access database, Excel, etc.

single row functions

can be applied to one or more columns in a record operate on a record by record basis all records are uniformly affected by single row functions

Append Query

combines tables together

fields tab

define how the data selected in the Fields panel should be used as part of a visualization

data model

define the relationships between tables

formatting tab

defines how the dashboard should look

Gestalt principles

describe how our mind organizes individual elements into groups to make sense of an entire visual

Pre-attentive attributes

emphasis quantity color

visual tools that aid in pre-attentive processing

emphasis, quantity, color

ERP Systems

enterprise resource planning SAP, Oracle, Microsoft Dynamics ERP, Oracle NETSUITE

Explanatory Data Visualization

explains what the audience needs to know shows specific relationships in data, such as link between causes and results usually used for client presentation

emphasis

form, position, motion

foreign key

has values consistent with the primary key of another table may have non-unique values across records

slicer/action/filters

help interlink multiple visualizations on a page

Tufte's Principles

highlight that "excellence in statistical graphics consists of complex ideas communicated with clarity, precision and efficiency"

analytics mindset

includes the ability to interpret and share the results of data analytics techniques with stakeholders

slicers

interactive visuals that enable the creation of focused data can choose any combination of values always interact with other visuals used to dynamically select a more focused data set use the Ctrl key to select multiple boxes

data transforming or cleansing

involves converting data from one format to another to load it into an analytics tool

primary key

is never blank or null, contains unique values across all records

top-down

leverage the objective/hypothesis to defile the analysis criteria know what you're looking for an you're going to find it

normalization

results in data being kept in different database tables to minimize storage cost and prevent duplication of records

'NUMBER OF ARTISTS'

measure that can be used in visualizations only

"Any" Data Type

mixes text and numbers

unstructured data

new form of data that is currently much more difficult to analyze, can't prove that this came from an automatic system

Edit queries

opens the query editor Apply Changes - save the changes you made and exit the query editor

dimensions

qualitative, categorical data, considered an independent variable e.g. the product (chair, table, desk, bookshelf)

flat file

raw data in its simplest form

data visualization

representing (often large) amounts of data in visual form so as to bring out hidden meanings, trends, and other attributes

stacked bar charts

use horizontal bars to make comparisons between categories, shows how different categories relate to one another using a numeric field this is represented as a bad, uses multiple fields to compare the categories

date masks

used to convert between date data stored as characters vs. the actual date data type

fields panel

used to select the data that will become part of the dashboard

Exploratory Data Visualization

used to understand the data, develop and assess a hypothesis or question or find a pattern in the data allows the audience to explore data for further analyses is conducted for a problem that has not been clearly defined

report view

workspace/toolbox for designing dashboards divided into 8 parts