The Life Cycle of a Data Project

Ever wondered how does a life cycle for a data project begin and end? Hint, its more of an ongoing process as In real-world scenarios, many of the tasks can be performed in a different order. The life cycle of a data project is in flux, meaning that steps could be backtracked or repeated depending on the results obtained.

Define the Business Problem

Easily considered the most critical step, in which the analyst identifies the business need and problems faced. This allows the analyst to consider the relevant data to be utilised for the business problem.

Studying and Cleaning of Data

Before running any algorithms or building models, the analyst should take the time to explore the dataset for patterns or what they can decipher from it. Afterwhich, cleaning of the data is essential if it contains issues such as missing data, extreme values or outliers. As some algorithms, e.g. the K-means clustering algorithm is susceptible to such problems, by cleaning the data, it allows a reduction on the adverse effects or results on the analysis performed. Data normalisation can also be applied to scale the data of an attribute into a smaller range.

Identify Data Analysis Techniques or Algorithms to be used