As enterprises increasingly strive to become data-driven operations that depend more on empirical evidence than...
on intuition to influence business decisions, they're going to need more people to be able to effectively crunch the numbers. Long gone are the days where the typical user is content waiting for a new report to be churned out by overworked BI analysts.
Analysis and presentation of numbers are expanding outside the reach of BI, as more organizations democratize the process. But in order to get the most out of what citizen data scientists can do across the enterprise, organizations must find ways to enable them to get the job done. Citizen data scientists need better access to not only the data that's relevant to them, but also user-friendly tools to analyze and present it in an understandable format.
Even though the business user is in charge of slicing and dicing, there's a lot of work on the back end to make that process as seamless and effective as possible. In particular, organizations seeking to enable citizen data scientists need to keep data visualization top of mind, as that's what will make or break the data-driven decision-making process.
"You need visualization to get the most out of big data," said Cody Swann, CEO of Gunner Technology, a development firm that aims to enable its users with visualization tools. "First, because visualization makes spotting potential trends and correlation much easier and, second, because seeing is believing, and execs need a visual representation to easily consume."
It takes a three-pronged partnership to enable a culture of self-service analytics so users can consistently, quickly and effectively build the right data visualizations when they need them. Users need not only support from IT to procure and run the infrastructure behind all the data, but also support from the BI analysts for whom data crunching is second nature, to help put the analysis and visualization process on rails for them. That way, users can then run individual reports on their own, but they're backstopped with the assurance that data handling is secure, runs consistently and is based on sound mathematical principles.
We asked experts for some tips to get organizations empowering their citizen data scientists. Here's what they suggested:
1. Pair citizen data scientists with business analysts
Self-service analytics served up to citizen data journalists shouldn't consist of just letting users figure it all out on their own. Ideally, organizations should pair the users with BI professionals to come up with different classes of analysis and visualization that work for everyone.
"Having a business power user paired up with a seasoned BI professional can often lead to better results faster than putting the responsibility into the hands of either group exclusively," said Michael Golub, senior vice president of analytics and machine learning at Anexinet. "Visualization is creative work just as much as it is technical. Taking an agile business/tech-working-together approach is one way to get the most out of the time spent designing and developing critical visualizations that will eventually help shape the direction of your organization or enterprise."
One way to divvy up the work is to have the BI team responsible for collecting data, aggregating it and setting calculations, said Lior Barak, consultant at Tale About Data.
"But the report building should be on the end user," he said. "He needs to use the KPIs [key performance indicators] supplied by BI and build the visualization answering his questions."
2. Pick the right self-service tooling
"We learned the hard way that rolling your own visualization tool is a fool's errand. It takes forever, costs a lot of money and [is] still not as good as the solutions that are out there," Swann explained. "You've got to go with a tool that makes visualization easy and flexible."
Samantha Marsh, marketing coordinator at iDashboards, said, as organizations seek to enable citizen data scientists, they need to look for self-service analytics tools with scalability and the ability to seamlessly integrate with data sources that matter to the business.
"You may also want to consider whether you'd like a templated or custom BI tool. If speed and simplicity is important to your organization, consider using a tool that kick-starts your projects with predetermined templates for visualizations," Marsh said. "Options like this can even come with prebuilt KPIs and dashboards. However, if you value flexibility and the ability to customize your visualizations, that may not be the right option for you."
3. Listen to the users
In order to get the tooling and processes that suit users, both IT and BI departments must listen to users.
"The hardest-learned lessons about big data and visualization is that only the end user knows how to define his needs, and no IT person can tell marketing people what he needs to see. We can consult them in the process, but to make decisions, they need to build the views that work best for them," Barak said.
Data scientists agree but emphasize that the citizen data scientists must put in the work themselves as well.
"Personally, I think tech teams should work closely with business users to understand the sort of visualizations they might want to produce to provide them the basic tools to start, but then business users should be investing time in learning these tools so they can work towards being self-sufficient in order to generate useful insights," said Joe Berry, data scientist at Edited.
4. Create personas
Obviously, a large organization can't tailor visualization processes for each and every citizen data scientist who needs to crunch numbers. In order to meet the widest number of users' needs, some experts recommended taking a page from the marketing department.
"Why not create personas for your visualization users, just like your marketing team creates them to market your products and services?" said Hugh Johnson, senior vice president of development at Suntico. "Then, when you are building your visualizations, understand for whom you are creating them and why. The result is more likely to be something that is genuinely useful for the business."
5. Vet the data
One of the big roles that IT plays in enabling citizen data scientists is in keeping the data feeds clean and well-handled. This includes ensuring that data is secure throughout the lifecycle and that users are given access to only the data they need to do their jobs.
"Typically, I would say that the IT function should be responsible for getting the data into a form that is ready to be visualized and analyzed and automating this process," Johnson said. "They should be responsible for making sure that the right controls are in place so that each visualization user can only see data that he or she is permitted to see."
Meanwhile, the BI function should also ensure that the data is truly vetted for the purpose for which it is intended.
"In many cases, the right data is not being collected at all. Without having data to work with, the analysts/front-end developers will have nothing to work with," said Justin Butlion, analytics and BI infrastructure specialist and founder of ProjectBI. "This needs to be the first thing that the head of BI focuses on, filling in these gaps."