A picture is worth a thousand words.
In fact, I’d argue we should say ten thousand, a hundred thousand, or even a million words. It’s just how we’re wired to process information.
There’s an oft-repeated “statistic” that states we process visuals 60,000 times faster than text. While there doesn’t seem to be an original source for that particular figure, everyone does agree that it is faster. We just don’t know by exactly how much.
- It takes only one-tenth of a second to comprehend a visual scene presented to us
- Roughly half of our brain is wired for visual input
- You process a symbol or pictograph and attach meaning to it in 250 milliseconds
If nothing else, you can “say” a lot more with a lot less if you use visuals rather than text alone.
Otherwise, we’d all be in danger of drowning in a torrent of text. Visuals eliminate the need for overwhelming numbers or complicated jargon.
And in the Big Data world we call home, that is very good to remember.
At its simplest, of course, Big Data refers to the very large data sets that are generated, collected, and often analyzed on an ongoing basis in the digital age.
And that vast array of data is a potential treasure trove for anyone willing to collect and analyze it… something possible for everyone in 2017 with a wide variety of “as-a-service” tools.
Once collected, analytic tools can sift through unimaginable mountains of data and identify patterns, trends, predictions, and insights that would take a human years to discover – if at all.
Those discoveries can help with everything from sales and marketing to customer profiles and customer service; from efficiency and productivity to education and healthcare; and on and on. The practical applications are many.
Big Data has volume, velocity, and variety. There’s a lot of it, it comes in fast and furious, and in many different forms:
- 1.7 megabytes of new information will be created every second for every human being on the planet by 2020
- The accumulated digital universe of data will grow from 4.4 zettabytes today to around 44 zettabytes, or 44 trillion gigabytes, in the next 3 years
- Every day, we create 2.5 quintillion bytes of data
But that’s part of the problem. The data – while incredibly useful – is just too much to wrap our heads around. It’s mind-boggling. We want to quickly understand what it is we’re looking at and what the data is saying.
The best way to do that? Visuals. Remember: a picture is worth a thousand words.
Big Data Visualization (or Dataviz)
Here are some good starting points if you’re wondering how to best display your data.
Let’s start with an easy one. Maps have been used to represent data visually for centuries. You can glance at a map and quickly understand – at least on a basic level – the physical makeup of a city, country, or the entire globe.
But they can get much more complex. Big Data could be presented on an interactive map – like this one that shows every dinosaur fossil ever found. Interactive maps allow the reader to click and drill down to multiple levels of detail.
A heat map can demonstrate the distribution of a particular metric (such as population density or number of cases of a virus) via intensity of color.
Maps, obviously, work best to represent data connected in some way to geography and location.
This one shows results of the 2016 US Presidential election by county, shaded according to winning candidate’s percentage of the vote.
Image Source: Wikipedia
The humble bar graph may be simple, but that’s kind of the point. We’ve been using them since grade school to easily represent simple data and comparisons. We can do the same with Big Data.
As you no doubt remember, a bar graph presents the value of variables based on the length or height of the bar. There is perhaps no better (i.e. easily digestible) way to compare two or more metrics, or to show change over time.
Bar graphs can be either single, grouped (aka clustered, representing the values of more than one item in each category), or stacked (aka composite, showing the proportions of the whole).
This format should be ranked by size (smallest to largest, or vice versa) most of the time, except when demonstrating change over time. You can choose to organize horizontally (best for comparative rankings) or vertically/by column (best for chronological data or comparing across different categories) depending on your data set and objective.
Image Source: Flickr
Another throwback to elementary math class, the line graph is still a mainstay of data visualization.
A basic line graph should be used to represent change over time. You plot individual data points over a period of time and connect them with straight lines. It doesn’t get any more straightforward than that.
The resulting line gives shape to the change taking place, and can be used to show the volatility, trend, acceleration (peaks), and deceleration (valleys) of your chosen metric.
You could also make use of the line graph to compare different items over the same period of time by using different colors for each item.
Just don’t crowd your graph with too many lines, or squish it with a value set on the axis that is much larger than the biggest data point.
A great example comes from this graph that shows the direct influence a hockey game had on water usage in Canada’s Edmonton during the Olympic Gold medal hockey game – with clear peaks as people flushed their toilets during the break, and troughs at important times in the game.
Image Source: Quora
Scatter Plot Graphs
These graphs require two sets of variables – an independent (x-axis) and a dependent (y-axis) – and whether there is a positive, negative, or zero correlation between them. Think of it as a line graph without the line.
Given enough data points (you need plenty), and the inclusion of a trend line, a scatter plot graph can display the existing correlation between the data points at a glance. Flower height and petal length, shoe size and IQ, age of driver and number of accidents, length of customer patronage and average order size… is there a correlation? A scatter plot graph can answer that question for you.
Take a look at this interactive chart which shows how GDP per capita correlates with life expectancy, and how that has changed over time.
Image Source: Plot.ly
Infographics are perhaps the most popular visual representation of data at the moment. They’re liked and shared 3x more on social media than any other kind of content, and 41.5% of marketers believe they have the highest engagement, with 30.4% listing infographics as the visual graphic they used most.
They’re appealing, they’re easy to understand, they’re attention-grabbing, they’re memorable, they’re easy to share, and they’re (relatively) easy to make. That’s a darn good list of attributes.
Anytime you have data ready to be shared, an infographic including pictographs, various charts, and limited text may be the fastest, easiest, most powerful way to get it to your readers. Represent your findings visually, link ideas logically, and tie it all together, and you create something they’ll gobble up and understand.
Simply put, infographics get results. Creating a compelling one is well worth your time and effort.
– An infographic by the team at Indusface
Okay, to be honest, a lot of people hate pie charts. With a passion.
“The pie chart is easily the worst way to convey information ever developed in the history of data visualization.” ~ Walter Hickey
Image via Chandoo.org
Not high praise. But to be fair, it’s all in how you use them.
The biggest strike against them is that people try to cram too much into them… and that’s not exactly the pie chart’s fault. Limit the number of slices, and it can be effective under the right circumstances.
And just what are the right circumstances? These:
- Each of the parts add up to a whole something
- Sufficient variation in the size of the slices (a slice with 24% and a slice with 26% are going to look pretty much identical)
- You’re only using ONE pie chart, and not comparing pie charts to other pie charts
Meet those criteria, and the pie chart may be your data visualization winner. As a rule of thumb, start with your largest slice at 12 o’clock, and work your way around.
Image Source: SlideTeam
Remember these from high school history?
They are simplicity itself, but timelines are a practical way to represent events over a selected period of time. Draw a line, mark the beginning and end, and add the major milestones/events between those two points. Done.
Whether it’s the history of your company, the history of your industry, your product catalog, or anything else you want to chart sequentially for your readers, make it into an interactive timeline and allow readers to click and get additional details for each point along it.
Fancy? No. But instantly recognizable and incredibly simple to comprehend. And isn’t that the whole idea behind data viz?
Image Source: Wikimedia Commons
Admittedly, this one is a bit of an oxymoron. How can a word cloud be an example of data visualization?
Short answer: it’s a collection of words and phrases in a more visual presentation.
A word cloud demonstrates how frequently a word (or phrase) appears in a block of text by connecting its size with its frequency. The bigger the word in the cloud, the more it appears in the text.
What is Donald Trump most concerned with in his latest speech? What topics or keywords appear most often in your social media account comments or product reviews? Use a word cloud tool and find out…and then share that answer with others.
The below word cloud was created from the text of the Declaration of Independence.
Image Source: AEA365
Anytime you want to painlessly explain a hierarchy, relationships, or a multi-step procedure, a tree diagram is probably the way to go (think of a family tree).
Readers can trace a particular process through all its possible manifestations depending on each step, or instantly see who reports to and supervises whom.
Tree diagrams pop up in science, math, genealogy, computer science, business, and more.
Image Source: Wikimedia Commons
There are many, many more types and styles to choose from depending on your data set and your particular objectives: icons, pictographs, proportional circles, chord diagrams, area charts, density plots, spider charts, and more. Compare a few formats to find the right visualization for your data.
It doesn’t matter how good the data you collect is if no one can understand it. High quality sources and analysis are wasted if the resulting insight is too complicated for everyone to look at and “get.” That’s where data visualization truly shines.
Visualization makes it easy to tell a story – and human beings love a good story – and make sense out of complex, boring, or extensive sets of data. It bridges the language and culture gap, becoming essentially universal to its audience.
So don’t just show the data. Tell a story with it.
A Few Weapons for Your Arsenal
There’s a tool or service available for almost everything in the modern world, and data visualization is no exception.
Here are a few tools we recommend to get you started:
- Silk – for interactive charts and maps
- Tableau – tools to help you create scatter plots, bar graphs, maps, and more
- Datawrapper – an open source tool; lets you easily create embeddable charts
- Chartio – combine data sources and execute queries from within the platform itself
- Timeline – generate beautiful interactive timelines
- Plot.ly – both 2-D and 3-D charts
- Chart.js – responsive, flat designs
- Exhibit – designed at MIT, Exhibit allows you to create interactive maps and other visualizations
- MyHeatMap – heatmap creation made easy
- Google Charts – a full set of data visualization tools from the good people at Google
Of course, data visualization obviously begins with data collection, and few tools can offer the kind of comprehensive web scraping you’ll find with Import.io. Sign up today for your free trial, and use the data you scrape with one of the tools listed above to see for yourself how easily and powerfully they can work together.
Finally, we’d be remiss if we didn’t mention David McCandless of Information is Beautiful. He offers workshops, an informative blog, and excellent examples for those new to the data visualization universe. Check out his TED talk here to get you started:
While you’re at it, head on over to our post 14 Fantastic Examples of Complex Data Visualized for even more inspiration.
Big Data isn’t going away anytime soon. If anything, it’s only going to get bigger. How you choose to use that data – for yourself, for your business, for your customers and readers – can make all the difference.