Join Doug Rose for an in-depth discussion in this video Create a question board, part of Learning Data Science: Using Agile Methodology.
- When you're working on your data science team, it's your research lead who's responsible for driving interesting questions. Coming up with good questions isn't an easy task. A good question can stir up a lot of new information. It forces people to rethink their work. That's why most organizations tend to shy away from good questions. A good question can cause some irritation. You're almost always itching to find the best answer. That can lead to a lot of work and sometimes even more questions.
In many organizations very few people tend to ask questions. In fact, usually only the best managers ask good questions. But even managers take a risk when asking questions. Too many questions are often seen as being not on the same page. Even today most organizations still try to focus on improving what they know. They figure that optimization will always put them ahead of their competitors. Good questions can often shake up well-ordered plans.
That's why they're usually left unasked. Good questions have a tendency to break up predictability. They can turn a well-ordered set of objectives into an open-ended exploration. It's the research lead's responsibility for breaking up this well-ordered process and injecting some exploration and experimentation. One of the best tools that a research can use for this is a question board. A question board is usually a large whiteboard filled with Post-it notes.
It's best placed near the data science team. There should be a lot of space with new questions and a short stack of Post-it notes in one of the corners. You might want to create a large arrow pointing down to the stack of sticky notes. Some teams will add the caption,"Ask a question." The whole point of the question board is to solicit questions from the rest the organization. The research lead drives the question. That doesn't mean that they come up with all the questions. There should be a combination of their own ideas, the data science team's questions and open-ended questions from the rest the organization.
The question board should be open and inviting. Try to make it look as enticing as possible by making it colorful and interesting. You'll want anyone to be able to walk by, grab a Post-it note and create a quick question. You may want to take special effort to keep it lighthearted and fun. Some teams even make it almost like a game. You can put a big bowl of candy next to the question board or they print out a sign that says, "Ask a question and win a prize." Another benefit of the question board is that it helps everyone else in the organization understand the purpose of the data science team.
The team will burn through questions and give short presentations. Often people recognize their own questions in the presentations. That way they'll be more likely to ask questions in the future. They may even encourage their co-workers to ask questions as well. You never run into the problem of too many questions. The research lead works with the rest the team to prioritize the most interesting ideas. If you get your organization to use the board then it starts to look a little bit like a three-dimensional search space.
You see the patterns in what people ask. The board itself becomes another data source. I once worked for an organization that put up a question board next to the data science team. At first it was just a curiosity. People would come by and read it the same way that people are drawn to an announcement board. The team was smart and they put it next to one of the water coolers. At first people just read it but didn't participate. After awhile a few questions popped up on the board. They were silly and didn't have much value. Still the research lead used the question board for communicating what the data science team was doing.
They posted the team's questions and continued to tell stories. Over the summer this organization brought in a whole new group of student interns. For the first month the students were trying to figure out the business. The students were more comfortable asking questions and the board soon filled up with their Post-it notes. Some of the questions they asked were very intuitive. They looked at the business from a fresh perspective. The questions were simple and well structured. The team started making them a priority and this helped explore the business in interesting new ways.
If you're the research lead try to put up a question board. It's a simple way to get interesting new questions while at the same time communicating your progress to the rest of the organization.
This course shows how to structure your work within a two-week sprint. See how to work within a data science life cycle (DSLC)—a methodology for cycling through questions, research, and reporting every two weeks. Explore key practices to help your team break down the work so it fits within a two-week sprint. Learn how to use tools like question boards to encourage discussion and find essential questions. And most importantly, learn how to grow your team's shared knowledge and avoid common pitfalls.
- Defining data science success
- Determining project challenges and criteria for success
- Using a DSLC
- Iterating through DSLC sprints
- Creating a question board
- Breaking down your work
- Adding to organizational knowledge
- Avoiding pitfalls