What is the data science methodology?

The data science methodology is a stepwise circular process used to answer a relevant question using data, techniques and technologies.

Step One: Business Understanding & Analytical Approach

Step one of the data science methodology involves developing a clear question and finding the appropriate analytical approach. This step can be likened to figuring out what type of meal you would like to cook for dinner and how you plan on realising the meal.

Step Two: Data Requirements & Data Collection

Step two of the data science methodology involves identifying and collecting the type of data needed to answer your clearly defined question. This step can be likened to buying, finding or growing the ingredients needed to make your meal.

Step Three: Data Understanding & Data Preparation

Step three of the data science methodology involves spending time understanding your data and preparing it for modelling. This step can be likened to washing and cutting the ingredients for your meal before the cooking.

Step Four: Modelling & Evaluation

Step four of the data science methodology involves building a working model (using the appropriate analytical approach) and evaluating its quality to ensure that your question is answered with a reasonable level of accuracy. This step can be likened to cooking and tasting your meal before eating.

Step Five: Deployment & Feedback

Step five of the data science methodology involves making your model available for others to use and receiving feedback to improve its ability to achieve the desired outcome. This step can be likened to eating your meal and sharing it with others gathering helpful feedback.