Data science is the field that uses statistics, programming and business knowledge to extract useful answers from data. The goal is to understand what happened, explain why it happened and predict what might happen.

How it works

The work follows a common path, which can be repeated several times:

  • Define the business question, such as "why do customers churn?".
  • Collect data from systems, spreadsheets and databases.
  • Clean and organize, which is usually the most time-consuming part.
  • Explore and analyze, with charts and statistics.
  • Create models, when it makes sense, and measure if they work.
  • Communicate the result clearly to decision makers.

Languages like Python and SQL are widely used in this process.

Practical example

A software subscription wants to reduce churn. The data person brings together usage and payment history, and compares who churned with who stayed.

They notice that customers who do not use a specific feature in the first weeks churn more. The company starts teaching that feature at the beginning, and tracks whether the churn rate changes.

SQL
SELECT recurso_usado, COUNT(*) AS clientes
FROM uso_primeira_semana
GROUP BY recurso_usado;

This simple query already helps raise hypotheses. The next step is to test whether the relationship is real and not a coincidence.

What it's for

Data science is for making decisions based on evidence. It shows up in sales forecasting, product recommendation, risk analysis, campaign optimization and quality control.

It also powers AI: machine learning models are trained with data that this area prepares and evaluates. Today, AI assistants help write queries and analysis code, but results need to be checked, because the model can be wrong.

Cautions

Correlation is not causation. Two numbers that rise together do not prove that one causes the other. Biased or incomplete data leads to wrong conclusions, and personal data requires care with LGPD.

Difference between data science and machine learning

Data science is the complete area, from the question to the decision. Machine learning is one of its tools, used when a model needs to learn patterns from data.