What Is Data Science?

A new way of turning data into decisions

Every time you stream a song, tap a map for directions, or get a fraud alert from your bank, a pipeline of data and algorithms is working quietly behind the scenes. Data science is the discipline that builds those pipelines: it combines statistics, computing, and subject-matter knowledge to pull reliable insight out of data and put it to use. A data scientist does not just compute averages; they frame a question, gather and clean the relevant data, explore it, model it, and communicate an answer that someone can act on.

What makes this moment different from a century of statistics is scale and speed. Organizations now record data on almost everything they do, cheap storage keeps it, and fast computers let us analyze it. The raw material is abundant; the scarce skill is knowing how to ask good questions of it and how to avoid fooling yourself with the answers. That skill is what this course builds.

The data science lifecycle

Real projects rarely run in a straight line, but they tend to move through a recognizable set of stages. Think of them as a loop you may travel more than once:

1. Ask a question. Turn a vague goal ("reduce customer churn") into something a dataset could actually answer.

2. Collect the data. Decide what to measure and how to obtain a sample that represents the population you care about.

3. Clean and wrangle. Fix missing values, inconsistent formats, and errors — usually the most time-consuming stage of all.

4. Explore and visualize. Summarize the data and plot it to see patterns, oddities, and relationships before committing to a model.

5. Model and infer. Use probability, statistics, and machine learning to estimate quantities, test ideas, or make predictions.

6. Communicate. Report the finding honestly, with its uncertainty, to the people who will make a decision.

A common rule of thumb is that collecting and cleaning data can consume the majority of a project's time. Beginners often imagine data science as clever modeling; practitioners know that the modeling only works when the earlier, humbler stages were done with care.

Three fields in one

Data science sits at the intersection of three areas. Statistics provides the language of uncertainty — how to sample, estimate, and avoid being misled by noise. Computing provides the tools to store, transform, and compute on data at scale, which is why Python is the working language of this course. And domain knowledge — understanding the medicine, the business, or the physics behind the numbers — is what separates a meaningful result from a meaningless correlation. Weakness in any one of the three tends to sink a project.

Here is the whole field in miniature. Suppose a streaming service wants to know which new users are likely to cancel. A data scientist would frame "likely to cancel" precisely, pull a sample of past users with their outcomes, clean the records, explore which behaviors track with cancellation, fit a model that scores each new user, and hand the marketing team a ranked list — along with an honest note about how often the model is wrong. Every lesson in this course develops one piece of that arc.

You do not need to be an expert programmer or statistician to begin. Comfort with basic algebra and a little Python will carry you a long way, and the ideas matter more than the syntax. Keep one question in mind throughout: not merely "what does the data say?" but "how much should I trust what it seems to say?"

Curriculum aligned with OpenStax's Principles of Data Science; all lesson text is original to Syllabus.