Zep - Core Principles For Data Visualization - Copy.pdf

Show the data People read graphs in a research report,article, or blog to understand the story beingtold.. But that does notmean that all of the data must be shown—indeed, many graphs sh

DATA VISUALIZATIONS

Show the data

People read graphs in a research report,article, or blog to understand the story beingtold The data is the most important part ofthe graph and should be presented in theclearest way possible But that does notmean that all of the data must be shown—indeed, many graphs show too much

REDUCE THE CLUTTER

Chart clutter, those unnecessary ordistracting visual elements, will tend toreduce effectiveness Clutter comes in theform of dark or heavy gridlines; unnecessarytick marks, labels, or text; unnecessary iconsor pictures; ornamental shading andgradients; and unnecessary dimensions Toooften graphs use textured or filled gradients

Standard research reports often suffer fromthe slideshow effect, in which the writernarrates the text elements that appear in thegraph A better model is one in whichvisualizations are constructed to complementthe text and at the same time to containenough information to stand alone As asimple example, legends that define orexplain a line, bar, or point are often placedfar from the content of the graph—off to theright or below the graph Integrated legends—right below the title, directly on the chart, orat the end of a line—are more accessible

Effective data visualization taps into thebrain’s preattentive visual processing.Because our eyes detect a limited set ofvisual characteristics (such as shape andcontrast), we combine variouscharacteristics of an object andunconsciously perceive them as comprisingan image Preattentive processing refers tothe cognitive operations that can beperformed prior to focusing attention on anyparticular region of an image In other words,it’s the stuff you notice right away

AUDIENCE

Always consider your audience —whetherthey need a short, written report, a more in-depth paper, or an online exploratory datatool

INCLUDE ANNONATIONS

Add explanatory text to help the readerunderstand how to read or use thevisualization (if necessary) and also to guidethem through the content

USE PIE CHARTS WITH CARE

We are not very good at discerning quantitiesfrom the slices of the pie chart Other charttypes—for example, bars, stacked bars,treemaps, or slope charts—may be a betterchoice

AVOID 3D

Using 3D when you don't have a third variablewill usually distort the perception of the dataand should thus be avoided

MAKE LABEL EASY TO READ

When applicable, rotate bar and columncharts to make the labels horizontal Ifpossible, make vertical axis labels horizontal,possibly below the title In general, makelabels clear, concise, and easy for yourreader to understand

TRY SMALL MULTIPLES

Breaking up a complicated chart into smallerchunks can be an effective way to visualizeyour data

USE MAP Carefully

Use maps carefully, always being sure it isthe geographic point you are trying to make.Column and bar charts, for example, areoften better at enabling comparisonsbetween geographic units

Color and font considerations

Avoid default colors and fonts—they all look thesame and don't stand out

Consider color blindness— about 10% of people(mostly men) have some form of color blindness

Avoid the rainbow color palette—it doesn't map toour number system and there is no logical ordering

Become a part of the team

at Zep

Why don't you start your journeyas a tech blogger and enjoyunlimited perks and cash prizes

every month.

Explore

grow & earn

zepanalytics.com