What is data science?

Data science is the practice of using data to answer questions, understand patterns and support decisions. It brings together statistics, computing and knowledge of the subject being studied. Machine learning can be part of that work, but data science also includes activities that need no AI at all: checking a spreadsheet, designing a comparison or explaining why a number cannot support a conclusion.

The work starts with a question. “What can we predict from these records?” is different from “Why are people waiting longer for appointments?” The second question may require interviews, a careful definition of waiting time and an examination of missing records before any model is useful.

A typical project involves obtaining appropriate data, checking what each field means, handling errors and missing values, analyzing patterns, and communicating what the results do and do not establish. A dataset is simply an organized collection of observations. Its columns may look tidy while hiding inconsistent definitions or changes in how information was collected.

Prediction and explanation also differ. A model may predict that a patient will miss an appointment without establishing why. An association between travel distance and non-attendance does not prove that changing distance alone would change attendance. Studying cause and effect requires additional assumptions or an appropriate study design.

Good data science makes its work inspectable. Another analyst should be able to see which records were included, how calculations were performed and where uncertainty enters. A clear chart that answers the actual question is often more useful than a complicated model whose output no one can act on.

In a referral service, a data-science project might find that most delays occur after a request for missing documents, not during clinical review. That finding could suggest changing ownership and follow-up rather than buying a more powerful AI model. Clinical and operational colleagues are needed to interpret the records: an empty field may mean “not done,” “not documented” or “recorded elsewhere.” Those are different problems, and treating them as the same can send an otherwise competent analysis in the wrong direction.

Sources: AMIA — Informatics: Research and Practice; Panch, Mattie and Celi — The “Inconvenient Truth” About AI in Healthcare (2019); Harvard Chan — Health Data Science

Further reading from Harvard Chan: Public Health Perspectives — To Meet Future Needs, Health Care Leaders Must Look at the Data (Science) (2022); Harvard Chan — Using advanced statistics and AI to improve health (2026).

To learn more, take a look at the AI in Health Care Certificate of Specialization from the Advanced Learning Academy at Harvard T.H. Chan School of Public Health. You can take an individual course or work toward the full certificate.

I’m a doctor, not a statistician. I teach the AI in Health Care Certificate of Specialization with Dr. Heather Mattie, a statistician and co-director of Harvard Chan’s Health Data Science master’s program. For more on data science and statistics, see Harvard Chan’s Health Data Science resources.

AI is a fast-moving field. This explanation reflects our class discussions and related material at the time of writing. Terminology, capabilities and tools may change.

For suggested additions or corrections, please email me at tpanch@hsph.harvard.edu.

Dr. Trishan Panch
Program Director, AI in Health Care
Harvard T.H. Chan School of Public Health — Advanced Learning Academy