Data Visualization Trends: Media and the Spread of Data Visualization
By Micah Melling, Chief Data Scientist, Americo Financial Life and Annuity
Micah Melling, Chief Data Scientist, Americo Financial Life and Annuity
A quick scan of leading news sites, such as The New York Times or the Washington Post, reveals an interesting shift in the way we receive information. Chances are the article you randomly select will feature a compelling visualization to illustrate a trend or expound a key development.
This is especially true with sports reporting. The website FiveThirtyEight, for example, was designed specifically to uses statistical analysis to tell stories about sports, politics, and other cultural trends.
In simple terms, a major trend in data visualization is for analysts to match the high expectations set by journalists. Much like the iPhone set expectations for intuitive user design, the media has established the new paradigms for effective data visualization. Our continual interaction with beautiful and powerful visualization has raised our expectations for such work. And this is especially prevalent in business; individuals interacting with data crave intuitive ways to consume it.
In many ways, data analysts act as designers, producing attractive visualizations that can be effortlessly understood.
"In many ways, data analysts act as designers, producing attractive visualizations that can be effortlessly understood"
Ultimately, the ever-increasing use of data visualization results in the democratization of technologies that produce powerful visualizations.
The spread of business intelligence (BI) tools is perhaps the prime example of this expanding access. Technologies such as Tableau, Power BI, and Qlik are commonplace in companies with any BI presence. Most users can produce effective visualizations through these platforms with little training or guidance; the front- and back-end software engineering is abstracted. With professional training or individual study, users can construct interactive and attractive graphics that would have involved a substantial time investment if manually coded.
Likewise, programming languages offer an increasing number of visualization packages. Python, which is gaining ground as the lingua franca of data analytics, has seen developers release multiple powerful visualization engines in recent years. Chief among them are Bokeh, Seaborn, and Altair. These libraries allow developers to produce an array of visualizations with a few lines of code.
Data visualization democratization is a boon for companies. However, this trend is not without concerns. For one, the accessibility of powerful tools can tempt us into becoming self-indulgent data visualizers. With a platform like Tableau, we can drag and drop features seamlessly and lose sight of interpretability. Likewise, analytics teams can inadvertently cause information overload.
The prodigious production of dashboards and visualizations can lead to audiences losing track of the most crucial information. Related, maintaining scores of dashboards is a gargantuan task, one that could lead to outdated dashboards being used for business decisions. An honest feedback loop along with a tiger-like focus on answering business questions, however, serve as the antidote.
Additionally, the ability to use intuitive tools can lull us into settling for passable work, not striving for outstanding. Even with a mighty BI tool or potent Python package, developing competent and streamlined visualizations that center on relevant business questions is no cakewalk. Nailing the nitty-gritty nuances that ward off misunderstandings is intensive work. While analysts do not have to author hundreds of lines of code to create a visualization, we are not free from a challenging task. Our duty pivots to one of a hybrid psychologist, communicator, and technical practitioner. Accepting this responsibility is crucial.
Lastly, the role of intellectual honesty and technical rigor cannot be forgotten. Improperly developed graphics can be unintentionally biased and misleading, creating more damage than benefit. Creating a culture of intellectual honesty and technical expertise is imperative. Not everyone who creates graphs needs to be an expert in statistics, though they should understand the dangers of manipulating axes and presenting sweeping findings from small sample sizes.
Data visualization for all is a reality, but the advantages are packaged with new business challenges. Those who ride the wave while proactively addressing the adjoined hurdles will be ripe for advancement.