Life Cycle of Data Science

Most Data Science projects, regardless of industry, follow a broadly similar life cycle - a repeatable sequence of stages that takes a business problem all the way to a working, monitored solution.

1. Problem Definition

Clearly identifying the business question - for example, "which customers are likely to cancel their subscription next month?"

2. Data Collection

Gathering relevant data from databases, APIs, logs, or external sources needed to address the problem.

3. Data Cleaning and Preparation

Handling missing values, duplicates, and formatting issues so the data is reliable enough to analyse.

4. Exploratory Data Analysis

Studying the data through summaries and visualisations to uncover patterns and generate hypotheses.

5. Modeling

Applying statistical or machine learning techniques to build a model that addresses the original question.

6. Evaluation

Testing the model on unseen data to check whether it performs well enough to be useful in practice.

7. Deployment and Monitoring

Putting the model into production, and continuing to monitor its performance since real-world data patterns can shift over time.

The life cycle is rarely a single straight line - it's common to loop back to earlier stages, such as returning to data cleaning after evaluation reveals unexpected model behaviour.

Coming Up Next

Next, you'll get an overview of the common tools used across the Data Science life cycle.