Data Visualization for Data Analysis introduces the analytical techniques and visualization tools used for examining data. It covers foundational design principles and data assessment, and the synthesis of technical analysis components into dashboards, presentations, and written reports. Students assess the ways in which visual representation assists the analysis and interpretation of complex data.
Within the SCNS taxonomy, CAP is the Computer Applications prefix, and the 3000-level number places this in the upper division — the junior year of a B.S. or B.A.S. in data analytics, information technology, or computer science. Polk State publishes it at 3 credits, 3 hours lecture. It appears at approximately three Florida institutions.
The distinguishing feature relative to the associate-level visualization course is that visualization here is treated as a method of analysis, not only a method of presentation. The framing — assessing how visual representation assists interpretation — points at exploratory data analysis, where the chart is how the analyst discovers what is in the data rather than how they report a conclusion already reached.
Florida carries visualization courses at both levels, and the distinction is the one Florida transfer actually enforces:
Lower-division credit generally cannot satisfy an upper-division requirement, so a student who completed CAP2741C will normally still need CAP3744 in a bachelor's program. The programming prerequisite is the other signal that these are different courses: an upper-division section is likely to use Python or R rather than a point-and-click tool. Read the syllabus — the toolset varies by institution and materially changes the work.
The idea that distinguishes an upper-division treatment. In exploratory use, the analyst makes many quick, ugly charts to find out what is in the data — distributions, outliers, missing patterns, relationships. In explanatory use, the analyst makes one careful chart to communicate a finding. These have different standards: exploratory plots should be fast and disposable, explanatory plots should be deliberate and polished, and confusing the two wastes time in both directions.
The practical consequence is that plotting the data first is a debugging technique. Anscombe's quartet is the standard demonstration — four datasets with identical means, variances, and correlation coefficients that look completely different when plotted. Summary statistics conceal structure; charts reveal it. An analyst who runs the regression before looking at the scatter plot is working backward.
The content most often skipped and most consequential. A point estimate drawn as a bar or a line looks certain, and readers treat it that way — including readers who know better. A forecast, a survey result, a small-sample average, and a model prediction all carry uncertainty that the default chart hides completely.
The available techniques — error bars, confidence bands, distributional displays, and ensemble or "hypothetical outcome" presentations — are all covered at this level, and the professional obligation is to use them. The failure mode is specific and common: a chart of a small subgroup shown at the same visual weight as a large one, which invites the reader to compare a number based on twelve observations with one based on twelve thousand.
The related judgment is aggregation: how data is binned, grouped, or smoothed changes what the chart appears to show, and there is rarely one correct choice. Documenting the choice is what makes the work honest.
A concrete methodological expectation at this level. Analysis performed by clicking is not reproducible — nobody, including the analyst, can reconstruct exactly what was done six months later, and the work cannot be rerun when the data updates or corrected when an error is found.
The practices that fix it: write code rather than manipulating files; keep raw data unmodified and derive everything from it; use notebooks or literate documents that combine code, output, and narrative; put the work under version control; and record data provenance — where each source came from and when it was pulled.
This matters professionally rather than academically. Analyses get challenged, findings get audited, and an analyst who can rerun their work and show the steps is credible in a way one who cannot is not.
Visualization is unusually effective at producing belief, and that effectiveness is exactly why the ethical content belongs in the course rather than as an afterthought.
Three specific obligations. Do not select the framing that produces the conclusion you wanted — choice of time window, baseline, grouping, and axis are all defensible individually and collectively make almost any story tellable. Show what the data does not support, including inconvenient subgroups and inconclusive results. And consider who is represented: aggregate visualizations of people can obscure disparities or, conversely, expose small groups to identification, which is a genuine privacy problem in health, education, and criminal justice data.
The professional habit that covers most of this: state the choices you made. A chart with a documented rationale for its binning, its window, and its exclusions is defensible; one without is not.
This course appears at roughly three institutions statewide, which is a small base. Content, credit value, and emphasis vary more than they would for a course taught at a dozen colleges, and one of the three may be a private institution rather than a Florida public college. Read your own institution's catalog description and syllabus rather than assuming this guide describes your section exactly, and have any transfer evaluated in writing.
The 3 credits and 45 contact hours reported here follow Polk State's published values — 3 credits, 3 hours lecture — with a programming prerequisite (COP3331, COP3835, or COP3177). Institutions offering it with a laboratory component will report more; confirm on your syllabus, and confirm the toolset, since a Python/R section and a Tableau/Power BI section are meaningfully different courses.
Assessment is typically project-based — an analysis carried from raw data to a report or dashboard — with critique of published visualizations as a recurring exercise. The critique work is more valuable than students expect: learning to say precisely why a chart fails is what makes your own charts better.
The first digit of an SCNS number denotes the year of offering, not transferability. Courses at the 1000 and 2000 levels transfer transparently between Florida public institutions, and 3000 to 4000 is unproblematic since both are upper division. The boundary that actually matters is 2000 to 3000, where lower-division credit generally cannot satisfy an upper-division requirement.
Generated September 2, 2026 · Updated September 2, 2026