Solving complex behavioral, business and computational problems require both a technical approach and equally a humanistic, artistic approach.
Like an artist, a data scientist must imagine and paint a picture. A picture that has meaning and can inspire their audience. They must see the picture in their mind first, and then painstakingly – line by line of code, dataset by dataset, build an image their audience can understand regardless of their audience’s technical skills.
And like an artist, if they find that the image they created, is not meaningful, not inspiring, and not going to serve their audience, they must be bold enough to throw out the canvas and start again.
A humanistic, artistic mindset matters because for most business problems the solution is not in the math, but the solutions lie in how well math is applied against a well-defined understanding of customer needs, behaviors and motivations. Human insight must guide ML/AI and statistical methodologies.
And like an artist, before any technical work is done planning and preparation come first. The canvas must be studied, the tools assembled, and the paints mixed. In data science that means:
- Clarifying and aligning on an opportunity/problem statement – determining what success looks before the work begins.
- Brainstorming, identifying, hypotheses, experiments and theories that can be challenged and explored.
- Gathering data sources and samples and understanding their potential utility in conducting experiments and testing hypotheses.
- Developing and documenting methodical workflows and exploratory techniques for planned experiments.