Starting with a question

"Always start with a question! Without a question to answer, there can be no meaningful analysis."

It's rare for us to start hunting for data without a question in mind. Even professional data analysts don't usually take a dataset and start looking for patterns without an idea of what they want to know first. This doesn't mean starting with a preconceived idea of what the answer to the question is, just that you have a starting point for inquiry.

A question defines the objective behind the 'wrangling' of the data.

The right question will help you to define:

• What is the story you have to tell?

• What is the problem you want to solve?

• Define the issue you are addressing

• What is the subject of your exploration?

• What is the topic of your conversation

Read more on the subject of questions

In the scientific method, the starting question is known as the hypothesis. For example, "Water boils at 60 degrees Celsius" is a hypothesis that can be proved right or wrong with an experiment. You can rephrase this as a yes or no question: "Does water boil at 60 degrees?" Journalists also start with hypothesis or question, such as "There was corruption in this procurement process" or "Which states recorded the highest proportion of votes for Peter Obi?". For more on the use of questions and hypothesis to lead investigative stories, we recommend reading the excellent (and free) Story-based Inquiry by Mark Lee Hunter.

Once you have framed your initial question, you can use it to inform where you should look for data. A question on inflation in Africa, for example, could take you to The World Bank or Africa Data Hub.