Creating Maps — Create symbol maps and filled maps using geographic roles

Historical Context & Motivation

The desire to represent data on maps predates modern computing by centuries. Early thematic cartography — the practice of overlaying statistical information on geographic boundaries — emerged in the early nineteenth century when cartographers began encoding demographic and economic variables into spatial representations. As datasets grew in size and complexity through the twentieth century, the manual process of plotting data points onto physical maps became untenable, creating a clear demand for automated geographic visualization tools. Tableau, first released in 2003 as a spin-off from Stanford University research on data visualization, addressed this need by abstracting away the complexities of coordinate systems and map projections, enabling analysts to produce sophisticated spatial visualizations with drag-and-drop interactions rather than programmatic GIS pipelines.

The central question that Tableau's mapping capabilities address is straightforward but powerful: how can we enable users who are not GIS specialists to produce spatially meaningful visualizations — symbol maps and filled maps — from ordinary tabular data containing geographic fields? The answer lies in Tableau's concept of geographic roles, which automatically geocode text fields like country names, state abbreviations, and ZIP codes into latitude-longitude pairs, bypassing the need for explicit spatial joins or coordinate lookups.

Core Principles & Definitions

Before constructing any map in Tableau, it is essential to understand the foundational abstractions that the platform uses to translate raw data into spatial visualizations. Tableau's mapping pipeline rests on a clear separation of concerns: geographic roles define what a field represents geographically, the mark type determines how data points render on the map, and the encoding channels (color, size, detail) control which dimensions of the data are visually communicated. Grasping these distinctions is analogous to understanding how a compiler separates lexical analysis from code generation — each stage has a well-defined responsibility.

Visual Explanation — Symbol Maps vs. Filled Maps

The diagram below illustrates the architectural difference between a symbol map and a filled map in Tableau. On the left, individual data points are plotted as circles at precise coordinates, with size encoding a quantitative measure (e.g., revenue) and color encoding a categorical dimension (e.g., product category). On the right, entire geographic polygons are shaded using a sequential color scale to represent an aggregated measure (e.g., total sales per state). Both map types share the same underlying geocoding pipeline but diverge at the mark-rendering stage.

Notice that the symbol map preserves the exact spatial position of each data point, making it suitable for datasets where precision matters — sensor locations, retail outlets, or event coordinates. The filled map, by contrast, abstracts individual records into region-level aggregates, which is appropriate when the analytical question concerns territorial comparisons: which states have the highest sales, or how does voter turnout vary across counties. In Tableau, switching between these two representations is often as simple as changing the mark type on the Marks card from Circle (or Shape) to Map, assuming the geographic role is correctly assigned.

How Tableau's Geocoding and Rendering Pipeline Works

Understanding how Tableau transforms a text column like State into a rendered map requires tracing the data through several stages: role assignment, geocoding lookup, coordinate generation, mark type selection, and tile-based rendering. This pipeline is conceptually similar to the stages of a compiler — each stage transforms the representation closer to the final output. Let's examine each stage in detail.

Stage 1 — Geographic Role Assignment

When Tableau connects to a data source, it performs automatic type inference on each column. If a column named Country or State is detected, Tableau auto-assigns a geographic role. Available roles include Country/Region, State/Province, County, City, ZIP Code/Postcode, Area Code, Airport, CBSA/MSA, Congressional District, and custom latitude/longitude. You can manually override the assigned role by right-clicking the field, selecting Geographic Role, and choosing from the menu. This is analogous to explicit type casting in a strongly typed language — you are telling the system how to interpret the data.

Stage 2 — Geocoding Lookup

Tableau maintains an embedded geocoding database (a SQLite file located in the Tableau installation directory) that maps each geographic name or code to a centroid latitude-longitude pair and, for polygon-based roles, to a set of boundary vertices. When you drag a geographic dimension onto the view, Tableau performs a join between your data values and this internal database. Any values that fail to match are flagged with a small indicator in the lower-right corner of the view, and you can resolve these ambiguities manually using the Edit Locations dialog. This is where hierarchical disambiguation matters: a city named "Springfield" exists in over 30 U.S. states, so Tableau uses the presence of a State or Country field in the view or data source to narrow the match.

Stage 3 — Generated Fields & Mark Rendering

Upon successful geocoding, Tableau creates two generated fields: Latitude (generated) and Longitude (generated). These are placed on the Rows and Columns shelves respectively, configuring the view as a scatter plot projected onto a map tile layer. The mark type on the Marks card then determines the rendering strategy. Setting it to Automatic or Circle produces a symbol map. Setting it to Map (the filled polygon type) triggers Tableau to fetch the boundary geometry from its geocoding database and render a filled choropleth.

Geographic Roles — A Detailed Classification

Tableau supports a rich taxonomy of geographic roles that govern how string and numeric fields are interpreted spatially. Each role maps to a specific level in a geographic hierarchy and determines whether Tableau can produce centroid-based symbol maps, boundary-based filled maps, or both. The table below enumerates the primary roles, their expected data formats, and which map types they support. Understanding this classification is essential because selecting an incorrect role is the most common source of map-rendering errors in Tableau.

A critical distinction in this taxonomy is between roles that support polygon geometries and those that only support point coordinates. The City and Airport roles, for instance, map to point centroids only, which means you cannot create a filled map from city data using Tableau's built-in geocoding alone. To produce filled maps for non-standard regions (e.g., sales territories, custom districts), you must supply your own spatial file (Shapefile, GeoJSON, or KML) and connect it as a spatial data source. This is the Tableau equivalent of implementing a custom data type when the standard library doesn't provide one.

Worked Example — Building Both Map Types from Scratch

Let's walk through a complete example using Tableau's built-in Sample - Superstore dataset. We will create two views: a symbol map showing individual order locations sized by profit, and a filled map showing total sales by state. This dataset contains fields for State, City, Postal Code, Sales, and Profit — all the ingredients needed for both map types.

Part A: Symbol Map — Orders by City, Sized by Profit

Part B: Filled Map — Total Sales by State

Symbol Maps vs. Filled Maps — Strengths & Limitations

Choosing between a symbol map and a filled map is not merely an aesthetic decision — it is a data-modeling decision with implications for accuracy, interpretability, and the types of questions your visualization can answer. The table below provides a structured comparison across several analytical dimensions that a computer science student should consider when designing geospatial visualizations.

Connections to Advanced Spatial Analytics

Symbol maps and filled maps represent the foundational layer of Tableau's spatial capabilities, but they connect directly to more advanced techniques that push the platform toward full GIS functionality. Understanding these connections prepares you to tackle complex geospatial projects and to evaluate when Tableau is sufficient versus when you should move to specialized tools like PostGIS, QGIS, or Python's GeoPandas library.

One particularly powerful advanced technique is the dual-axis map, which combines both map types in a single view. By duplicating the Latitude field, placing it on both the Rows shelf slots, and synchronizing the axes, you can render filled state polygons on one axis and overlay proportional circles on the other. This composite visualization answers questions that neither map type can address alone — for example, showing regional sales volume (filled) alongside individual store profitability (symbols). In Tableau's architecture, dual-axis maps exploit the same mechanism as dual-axis charts in non-spatial contexts, treating latitude as a continuous measure that can be independently configured per axis.

Practice Problems

Lesson Summary