Albert Einstein has been credited with saying that if he had an hour to solve a problem he’d spend the first 55 minutes analyzing the problem and the last five minutes solving it. Whether Einstein actually said this is subject to debate, but it’s a good quote and very relevant to the topic of data science. However, in the context of data science, I would reword it as “If I had an hour to solve a problem, I’d spend the first 55 minutes asking questions.”
This statement is true even outside data science in our daily interactions with people and problems. Just think of how many times you became upset over an issue unnecessarily because you didn’t fully understand what was going on. How many times have you argued with a loved one or a colleague only to realize later that the argument was based on a misunderstanding that could have been avoided simply by asking one another the right questions first? How many times have you tried to fix something before you knew the cause of the problem, only to make it worse through your efforts?
Questions are the key to discovery and learning, which is why asking questions is such an important part of what a data science team does. You can have the best data warehousing and business intelligence (BI) tools on the market and total access to huge data sets, but if your team struggles to ask compelling questions, the technology and data are virtually useless.
It’s a Cultural Issue
For most organizations, asking compelling questions is not an easy task. Their corporate culture isn’t geared for asking questions and analyzing data. It’s geared for setting goals and objectives and implementing strategies and tactics to achieve those goals and objectives. Some companies even discourage management and employees from asking questions, because questions are seen as inhibiting action and progress. There’s even a term for it: “paralysis by analysis.”
People are expected to attend meetings in which they’re told what to do. When they leave a meeting, they’re expected to implement the plan. That’s why organizations place so much value on experts. Experts deliver answers, which are final and solve problems. Questions are ongoing and often reveal problems that nobody wants to address. However, asking questions and uncovering problems are what opportunities and innovation are all about.
The Research Lead
A good research lead on a data science team has a knack for asking compelling questions. For example, have you ever waited in a long line to board an airplane and wondered, “There has got to be a better way.”? If so, you might make a good research lead. You might ask the question, “Is there a better way to board an airplane?” I have asked this question many times myself. My first thought was that people should board from the back of the plane to the front, so the aisles aren’t cluttered with people trying to stuff their carry-ons in the overhead compartments. Then I wondered why airlines aren’t doing this already; it seems too obvious a solution to be overlooked. So, I asked another question. I actually asked Google, “Why don’t airlines board passengers starting at the back of the plane?” There’s actually a good reason. The three wheels that support the plane are positioned toward the front; loading everyone from the rear forward could potentially tip the plane back.
Astrophysicist and frequent flyer Dr. Jason Steffen came up with another solution. He imagined people boarding in parallel from back to front, starting with passengers in the window seats, and skipping every other row. The process would enable more passengers to be working at the same time loading their carry-on bags into the overhead storage and taking their seats, instead of following a serial process in which passengers have to wait for those in front of them. (Although this method would be much more efficient, it hasn’t been adopted for a variety of reasons, including the fact that people flying together would need to board separately.)
Dr. Steffen embodies three essential qualities of a good research lead: 1) A logical mind. 2) Some knowledge of the business. 3) The ability and drive to question assumptions. While a large majority of people accept the standard airline boarding routine, Dr. Steffen was bold enough to ask, “Is there a better way.”
If your organization struggles to ask questions, leadership may need to take steps to change the corporate culture. Curious, innovative personnel should be rewarded, not stifled. They should be given the time and freedom to explore the data and come up with new ideas. Every organization should also have a data science team in place focused exclusively on mining data to extract knowledge and insights. The research lead on this team should be someone who knows the business and has a knack for asking compelling questions.
Frequently Asked Questions
Why is asking the right questions so important in data science?
Asking the right questions is very important in data science because it helps us understand the problem better. If we don't ask good questions, even the best tools and data won't help us much. Good questions guide us to the right answers and help us find useful information.
What are some common barriers organizations face in fostering a culture of asking questions?
Organizations often face several challenges in encouraging people to ask questions:
- Company Culture: Many companies focus more on setting goals and getting things done rather than asking questions and analyzing data.
- Discouragement: Some companies discourage asking questions because they think it slows down progress.
- Meeting Structure: In many meetings, people are told what to do instead of being encouraged to ask questions and discuss ideas.
How can organizations change their culture to encourage more questioning and innovation?
Organizations can do several things to encourage more questioning and innovation:
- Reward Curiosity: Leaders should reward employees who are curious and come up with new ideas.
- Change Meetings: Meetings should be designed to encourage asking questions and discussing ideas, not just giving orders.
- Focus on Data Teams: Companies should have special teams that focus on analyzing data and finding new insights. These teams should include people who are good at asking important questions.
What qualities make a good research lead in a data science team?
A good research lead in a data science team should have:
- Logical Thinking: The ability to think clearly and logically about problems.
- Business Knowledge: A good understanding of the business and industry they are working in.
- Questioning Mindset: The drive to question existing ideas and look for better solutions. For example, Dr. Jason Steffen questioned the usual way of boarding airplanes and came up with a new idea.
How can data scientists balance the need for thorough questioning with the pressure to deliver quick results?
Data scientists can balance asking thorough questions and delivering quick results by:
- Prioritizing Questions: Focus on the most important questions that will have the biggest impact.
- Iterative Approach: Use an iterative process where they start with quick insights and then refine them over time.
- Clear Communication: Explain to stakeholders why asking thorough questions is important and how it leads to better solutions in the long run.
This is my weekly newsletter that I call The Deep End because I want to go deeper than results you’ll see from searches or AI, incorporating insights from the history of data and data science. Each week I’ll go deep to explain a topic that’s relevant to people who work with technology. I’ll be posting about artificial intelligence, data science, and data ethics.
This newsletter is 100% human written 💪 (* aside from a quick run through grammar and spell check).
More sources
- https://towardsdatascience.com/how-to-ask-the-right-questions-as-a-data-scientist-913621907411
- https://hbr.org/2016/11/better-questions-to-ask-your-data-scientists
- https://www.forbes.com/sites/kalevleetaru/2019/06/17/sometimes-data-science-is-not-about-giving-answers-its-about-asking-better-questions/
- https://towardsdatascience.com/how-do-data-scientists-ask-the-right-questions-6a5d7b89cdd9