Ganes Kesari has a secret that he wants you to know: storytelling is vital to successfully presenting data analytics. This was the focus of his talk on Thursday as part of the Center for Data, Mathematical & Computational Sciences Lecture Series.
Kesari, a co-founder of the data science company Gramener and a past TEDx speaker, started the talk by presenting examples of failed data visualizations. From percentages not adding up in a CNN graphic to the size of bars on a chart not matching the numbers during a broadcast news weather report, the wider the reach, the more embarrassing the mistake.
Kesari emphasized that stories are impactful because they carry emotion and stick in people’s memory. “We remember the stories we heard in childhood,” he said, “That’s why we can create a data story. Perhaps it will be remembered even beyond the last meeting you attended.”
The purpose of data storytelling is to create a “compelling narrative based on data analytics to influence and inform” the audience in a simple and easily understood way. The four aspects to consider are the audience, data, visual design and narrative. It’s important to start by considering who the target audience is because every person looks at the data with different needs and concerns.
“Unless you know which questions you’re going to be answering with analytics, with visualization, how can we pick the right visual or which answer is important?” Kesari said.
He then moved on to explain the importance of gathering the correct data and choosing the right analytics approach. The three main types of analytics approaches are descriptive, diagnostic and predictive, which all help a person discern what their focus should be and what kind of questions they should ask about the data.
“In the real world, data sets are dirty and dangerous,” Kesari said with a chuckle. “You need to transform the data into a format that you can work on.”
Kesari used an interactive example to show how effective visualization can be. He asked the audience to find certain groups of numbers first on blank sets and then on sets with those certain groups highlighted in green, which made the process much more efficient.
This example also demonstrated how easy it is to show patterns in data. Kesari emphasized how knowing the pattern that you want to convey helps inform what style of visualization is the most effective.
“When you identify the question, you’ve identified the insight, and you know what kind of pattern you want to show, it becomes easier,” he said. “This is a cheat sheet.”
The final step is to build an emotional arc that keeps the listener engaged with the data you’re sharing. Kesari recommended two ways of doing this: using a conventional structure that starts with the context and leads to a conclusion, or an inverted pyramid structure that starts with the takeaway and ends with the context. He also suggests letting the format of the presentation inform the way you structure the narrative.
Above all, Kesari’s point was that data visualization is a powerful tool that often goes unnoticed, but one presenters can use to their advantage when sharing data analytics.
“You can have the greatest idea. If you are not able to convince enough people, it doesn’t matter,” he said. “That’s why you need data visualization.”
rgatherc@ramapo.edu
Photo by Rebecca Gathercole
