Data science uses data to investigate questions and support decisions. The work includes deciding what to measure, writing code, checking data quality, and explaining what an analysis can support.
A degree can combine computing, statistics, mathematics, and study in another subject. Read the requirements at each university. Programs with the same name can have different prerequisites and emphases.
What will you learn?
Programming helps you work with data and repeat an analysis. Statistics helps you reason about variation, uncertainty, and the strength of evidence. Mathematics supports the methods behind that work.
Subject knowledge matters too. A model of environmental measurements needs an understanding of those measurements. Code that runs successfully can still answer the wrong question.
UC Berkeley's Data Science major page gives one example of a degree structure. Its Data 8 course introduces computational and statistical reasoning. These are examples, not a universal curriculum.
A first project you can finish
Choose a small public dataset from the student datasets guide. Start with one question, such as how a recorded measurement varies across a selected period.
- Read the documentation and identify what one row represents.
- Write down the units and missing-value codes.
- Select a small, justified subset of the data.
- Make one graph with labeled axes.
- Explain the pattern and one limit of the comparison.
Use a spreadsheet if it fits the task. Learning a new programming library is not a prerequisite for every useful analysis.
Keep the original download and record every change. A reader needs to know which rows you removed, not just see the final graph.
Prediction and explanation are different tasks
A prediction model estimates an outcome from available information. A model that predicts well does not automatically explain why an outcome occurs.
For a prediction project, keep evaluation data separate from the data used to fit the model. Record the baseline you compare against. Otherwise, a complicated model can look impressive without improving the result.
For an explanatory claim, examine the study design and alternative explanations. An association in public data does not establish a cause by itself.
Prepare in high school
Statistics and computer science are useful places to start where available. Practice reading graphs, writing short programs, and explaining why a method fits a question.
You can also use the research-project planner with a school club or a personal project. A finished, documented analysis gives you something concrete to discuss with a teacher.
Explore the computer science opportunity collection for courses, camps, and research routes. Check prerequisites before assuming a program is intended for beginners.
Questions to ask a university
Ask which mathematics and programming courses precede the major. Find out whether students study data ethics, communication, and a subject outside computing.
Ask to see examples of undergraduate projects. Notice whether students explain limitations, cite datasets, and evaluate their methods. Those habits matter more than a long list of tools.
Report your work
Use the STEM-paper guide to explain the question, data, method, and result. Share the source and analysis steps with your project so another student can follow them.
Do not include private records in a public notebook. Public access to a tool does not grant permission to publish someone else's information.
