Feverish coding and complex statistical techniques to generate data is just one aspect of a data scientist’s job, but constructing a narrative and communicating results to be accessible to non-technical audiences is just as important. In this year’s Plug in to Data Science (Nov. 7–8), keynote speaker Brian Xu, a data scientist at LinkedIn, will discuss the crafting of a data story through the lens of a LinkedIn analysis evaluating the impact of the recent U.S. government shutdown.
Xu’s keynote presentation, “Crafting a Data Story: U.S. Government Shutdown Impact,” will take place on Nov. 7 at 12:30 p.m. CDT. We spoke to him about the current challenges of storytelling through data science, as well as how these narratives tie back in to the work of prospect development professionals.
How do you believe the ways in which authors write and frame their articles affect the public’s perception of complex issues?
Journalists have the ability to shape how their audience perceives a topic, and their points of view can differ, even when they look at the same data set. One writer could take certain pieces of a dataset and claim that “the economy is slowing down.” Another writer could use other parts of the data set to say that “the economy has grown since last year.” Further, both could be true if there were measures indicating that the economy was still growing but that the growth had slowed recently. My role as a data scientist includes ensuring that there is accurate data on a topic that is interpreted and represented correctly.
When data scientists prepare to write a narrative, what are the most common challenges they come across? What are some of the tools they can use to overcome these challenges?
When building a narrative, a common challenge I’ve come across is realizing that the data table or chart I have generated is not the finish line. I quickly learned to take a step back and ask myself simple — but important — questions, such as what does this data mean? Why does this matter? Communicating findings clearly and doing more research to add context is outside the comfort of code and numbers, but it’s often essential. A strong story either starts a conversation or adds to a conversation that is already happening, and typically includes elements like human stories and experiences to support data.
How do you believe the crafting of data-driven narratives has changed over time, and where is there room for improvement in the future? How does this translate to the work prospect development professionals do?
Two things stick out to me. First, there is more interest in stories that are grounded in data — especially at a time when there’s so much information competing for our attention. And second, we have access to more data now than ever before, which has opened up a much more complex, layered category of insights we can create. The combination of these two forces means we need to train more data scientists with soft skills that help them generate and find the data points that fuel data-driven narratives. Data-driven storytelling is also immensely important for prospect development professionals, who use data to identify trends and develop compelling narratives to reach decision makers.
There is more interest in stories that are grounded in data — especially at a time when there’s so much information competing for our attention.
How does LinkedIn’s role as a combined news medium and networking platform affect the process of creating and distributing data-infused news?
At LinkedIn, our members’ actions and profiles help us uncover the latest trends in the world of work, primarily around three key areas: hiring, mobility and skills. LinkedIn’s team of editors write stories that engage members in conversation on their feeds, and we also partner with organizations like the World Bank and World Economic Forum to help shape their research on future of work-related topics. You can follow my LinkedIn profile to check out the latest stories and content I’ve been working on and the LinkedIn Editors page to keep up with all the latest stories from the team.
What would you like attendees to take away from your session?
I hope that other data scientists are able to learn from how we’re approaching data-driven storytelling at LinkedIn and apply some of the key lessons I’ve learned to their own work.
Brian Xu is a data scientist at LinkedIn working to turn LinkedIn’s Economic Graph data of more than 645+ million members into insights and stories that help create economic opportunity for every member of the global workforce. Brian produces LinkedIn’s official monthly Workforce Report, which shares hiring, migration, and skills gaps insights and is covered by prominent media outlets in conversations on the labor market and economy. Brian has also worked on numerous projects with partners like the World Bank seeking to make LinkedIn data available to help policymakers around the world understand their workforces and make decisions. For more information on the LinkedIn Economic Graph, please visit economicgraph.linkedin.com
This article relates to the Data Science domain in the Apra Body of Knowledge.
For more Data Science content, check out these bundles: