Data scientists turn messy data into decisions — cleaning and preparing it, building and validating statistical and machine-learning models, and explaining the result to people who will act on it. ⚠ There is no licence, no entrance examination and, for most people now working in the field, no data science degree: the occupation is older than the major.
Licensure and accreditation
There is no licence and no gatekeeper, which cuts both ways. Nothing stops you entering from any field — and nothing certifies that you can do the work, so employers judge demonstrated skill instead of credentials. ⚠⚠ WATCH THE DEGREE LEVEL. O*NET's survey respondents split 48% bachelor's degree required against 44% MASTER'S, which is close to an even division and far higher than mechanical engineering (52% bachelor's) or electrical engineering (82%). Plan on the likelihood that a master's is expected for the research-heavy roles, while analyst and data-engineering roles are reachable with a bachelor's and a portfolio. ⚠ Certificates from cloud vendors and course platforms are common and are worth something for specific tools; they do not substitute for the statistics.
⚠⚠⚠ There is no entrance data for this path, and that is the finding
On the Lawyer page this section reports exactly what law-school applicants majored in, because the Law School Admission Council counts every applicant and publishes it. Nothing equivalent exists for data science, because nothing equivalent can: there is no admissions body, no licensing board and no examination. Anyone who tells you that a particular major is the route into data science is guessing.
And the degree is newer than the job. Florida’s public institutions awarded 84 data science credentials in 2023 across ten schools, and 104 in data analytics. For comparison, in the same year they awarded 6,122 in computer and information sciences, 1,111 in economics, 735 in management information systems and 467 in mathematics. Practically nobody now working as a data scientist holds a data science degree — they could not have.
So the fields listed below are not a survey of where data scientists came from. They are the Florida programmes whose curricula actually deliver the skills the work requires, with the number of graduates each produces so you can see the scale. That is a weaker claim than the Lawyer page makes, and it is the honest one.
Projected need
O*NET, on federal projections: 245,900 data scientists employed (2024), growth projected “much faster than average” (7% or higher) through 2034, and about 23,400 openings a year. Median pay $120,230 ($57.80 an hour), 2025.
⚠ Note the ratio: 23,400 annual openings against 245,900 employed is a far higher churn-and-growth rate than mechanical engineering (18,100 against 293,100) or electrical (11,700 against 192,000). It is a genuinely expanding field — and one where job titles shift under you, so read the duties rather than the title.
Necessary skills
O*NET lists the tools as Python, R, Java and Scala; SQL and databases including PostgreSQL and MongoDB; AWS and Azure; SAS, TensorFlow and MATLAB; Power BI, Tableau and Looker. Tools change; the four things underneath them do not:
- Statistics you actually understand. ⚠ This is the part people skip and the part that separates a data scientist from someone who can call a library. If you cannot say what a p-value is the probability of, or why a model that fits the training data beautifully may be worthless, the rest is decoration.
- Programming that survives contact with real data. Not a first course — the ability to write code someone else can read, test and rerun. Python and SQL are the working minimum.
- Data wrangling. Practitioners consistently report that most of the job is finding, cleaning and reconciling data. Coursework rarely reflects that ratio; a real project will.
- Communication. The deliverable is a decision someone else makes. A correct analysis nobody acts on has failed.
- Domain understanding. ⚠ This is why entering from economics, biology, psychology or engineering is a genuine advantage rather than a handicap — you already know what the numbers mean.
⚠ The mistake to avoid
The visible part of data science is modelling, so the temptation is to collect machine-learning courses and certificates and skip the statistics and the data handling. That produces someone who can fit a model and cannot tell whether it means anything — the most common failure employers describe. Take the statistics sequence properly, and prefer one substantial project you can explain end to end over five short courses you cannot.
Advice
- Choose a major you will do well in, then add the quantitative spine. Statistics, programming, databases and linear algebra can be added to almost any degree; genuine domain knowledge is much harder to bolt on afterwards.
- Build a portfolio, because there is no credential. Two or three projects using real, messy public data — documented, reproducible, with the reasoning visible — do more than any certificate. This field hires on evidence.
- Plan for the master’s question early. With 44% of respondents reporting a master’s requirement, check the job adverts for the roles you actually want before you finish year two.
- If you are starting at a Florida state college, the associate-level data science and data analytics programmes exist at ten institutions each and articulate into baccalaureate work — but confirm the receiving university’s requirements in writing, because this is a new and fast-moving area where transfer agreements lag.
- Learn SQL earlier than feels necessary. It is the least glamorous item on the list and the one used every day.
There is no single required programme for this path. These are the fields students
actually arrive from — each is a place in the catalog you can browse.
A program is a field of study, one level above a degree — each page says which degrees it
covers and which Florida institutions award them. ⚠ The count is institutions that
awarded a credential in the field in 2023;
it is not a promise that a school runs this exact degree, and it counts something different
from the course-based list further down.
Programs worth knowing about if you are considering this career — ⚠ they do
not lead to it. Each row says what it is and is not.
Selected, not exhaustive — the courses specific to this path plus the foundations it is built on.
A course number does not identify a course in Florida: always check your own
institution's catalog and take a syllabus to an adviser.
STA2023
—Elementary Statistics
Guide
Statistics — the foundation of everything on this page, and the course whose absence employers notice. Take it seriously rather than clearing it.
MAC2311
—Calculus I
Guide
Calculus I — required by the statistics and machine-learning sequences that follow, and by most quantitative majors.
MAC2312
—Calculus with Analytic Geometry II
Guide
Calculus II — integration and series, used throughout probability theory.
MAD2104
—Discrete Mathematics
Guide
Discrete Mathematics — logic, sets, combinatorics and graphs: the mathematics of algorithms and of relational data.
COP2071C
—SQL Fundamentals
Guide
Querying Databases with SQL — the least glamorous skill on the path and the one used every single day. Learn it earlier than feels necessary.
COP2700
—Database Design and Analysis
Guide
Introduction to Database — relational design and modelling; why data is shaped the way you find it.
COP3530C
—Data Structures and Algorithm Analysis
Guide
Data Structures — how to handle data at a scale where the naive approach stops working. The dividing line between scripting and engineering.
STA4321
—(gm)probability and Statistics
Guide
Introduction to Mathematical Statistics — probability and inference derived rather than recited. ⚠ This is the course that makes you able to judge a model instead of just fitting one.
STA4234
—Introduction to Regression Analysis
Guide
Regression Analysis — the workhorse method of applied data science, and the one most often misused.
STA4024
—Statistics II
Guide
Statistics II — the applied second course: inference, ANOVA and modelling beyond the introduction.
STA4222
—Sample Survey Design
Guide
Sample Survey Design — how data is gathered, and therefore what it can and cannot support. Underrated for anyone who will work with survey or observational data.
CAP4770
—Data Mining
Guide
Data Mining — pattern discovery, classification and clustering; the closest undergraduate course to the modelling half of the job.
CAP4786
—Introduction to Big Data Analytics
Guide
Big Data Analytics — distributed processing and the tooling for data that will not fit on one machine.
CAP4774
—Data Warehousing
Guide
Data Warehousing — how analytical data is organised in practice, which is where most working data actually lives.
COP4710
—Database Systems
Guide
Database Systems — the internals: indexing, query planning and transactions. Why a query is slow, and what to do about it.
CDA3101
—Introduction to Computer Organization
Guide
Computer Organization — what the machine is doing underneath, which is what makes performance problems tractable rather than mysterious.
ENC1101
—English Composition I
Guide
Composition I — the deliverable is a decision someone else makes on your explanation. Writing is not a soft extra in this field.
39 Florida public institutions teach courses on this path. The count is
how many of the 17 listed courses each one carries — it shows where
the path is best covered. It is not a statement that a school offers this as a
degree; for that, check the institution’s own programme list.
Requirements change. These are the authorities that decide them — check them directly before
committing to a route.