09 · Maps & Geographic Data

Tableau recognizes common geographic fields automatically and can plot them on a built-in map with no extra setup. This module extends Orders with a State column and builds a filled map of Sales by state.

1. Extending the dataset with geography

For this module, assume the Orders table (Module 1) has one more column, State, assigning each order to a US state within its region:

Order ID	Region	State	Sales
1001	East	New York	1200
1002	West	California	450
1003	East	New York	60
1004	Central	Texas	800
1005	West	California	2200
1006	East	Massachusetts	950
1007	Central	Texas	120
1008	West	Washington	1100

2. Geographic roles

1. In the Data pane, a field named "State" (or "Country", "City", "Zip Code", and similar common names) automatically gets a small globe icon, meaning Tableau assigned it a Geographic Role — here, State/Province.
2. If a field isn't auto-detected (e.g. it's named Location instead of State), right-click it → Geographic Role → pick the correct role manually (State/Province, Country/Region, City, ZIP Code, etc.) so Tableau knows how to map its values to latitude/longitude.
3. Misspelled or ambiguous values (e.g. "Calif." instead of "California") show up as an unrecognized exclamation-mark icon on the field or a small "N unmatched" indicator — click it to manually match ambiguous values to the correct real-world location.

3. Building a filled map

1. Double-click State in the Data pane (or drag it to the canvas directly) — Tableau auto-generates Longitude and Latitude fields onto Columns and Rows and switches the mark type to Map, with one dot per state.
2. Drag Sales onto Color on the Marks card, and change the mark type dropdown from Automatic/Circle to Map → Filled Map (or use Show Me and pick the filled-map thumbnail) so each state renders as a shaded polygon rather than a dot.
3. Verify by hand: California's total Sales = 450 + 2200 = 2650 (highest, darkest shade), New York's = 1200 + 60 = 1260, Texas's = 800 + 120 = 920, Massachusetts's = 950, Washington's = 1100 — California should render as the darkest state on the color legend's high end.

4. Dual-axis maps (points over a filled background)

1. A common pattern layers detail points on top of a filled region map: duplicate the Latitude field (drag a second copy of Latitude onto Rows, next to the first), giving two overlapping map layers you can style independently.
2. Right-click the second Latitude pill → Dual Axis. Set the first layer's mark type to Filled Map (colored by, say, Region) and the second to Circle sized/colored by Sales — producing a background of regional shading with individual order markers on top.
3. This layered technique is the standard way to combine a "where" (filled regions) and a "how much, precisely, at this point" (sized circles) story in one map.

5. Limitations to know

1. Tableau's built-in geocoding covers standard administrative levels (country, state, county, city, ZIP) — a custom territory (e.g. an internal "Sales Territory 7") needs either a custom geocoding import (Map menu → Geocoding → Import Custom Geocoding) or joining to a reference table that maps it to a recognized geography.
2. A filled map's shading always represents an aggregate over the whole region (e.g. a whole state) — it cannot show sub-state variation without a finer geographic field (county, ZIP) actually present in the data.

How It Actually Works

A Tableau map is not a special rendering surface — it's the same GROUP BY-and-aggregate query as any other chart, plus a geocoding lookup table (bundled with Tableau, mapping recognized place names to latitude/longitude polygons) that VizQL joins in behind the scenes:

1. Double-clicking State generates a query like SELECT State, SUM(Sales) FROM Orders GROUP BY State, exactly like a bar chart would — but because State carries a Geographic Role, Tableau also performs an implicit join against its internal geocoding table on the State value, attaching a polygon boundary (for a filled/shaded map) or a latitude/longitude point (for a symbol map) to each returned row. An "unmatched values" indicator (Section 2.3) is literally a failed join — a State value with no matching row in the geocoding table.
2. A filled map's shading is still just SUM(Sales) per state from that same GROUP BY — verify: California = 450 + 2200 = 2650, New York = 1200
3. 60 = 1260, Texas = 800 + 120 = 920, Massachusetts = 950, Washington =
4. California, at 2650, should render as the darkest polygon — the color encoding is a post-query rendering step (a color scale mapped over the aggregate value), not a different query.
5. A dual-axis layered map (Section 4) runs as two separate queries sharing the same geographic axis — one per axis/mark layer — which is exactly analogous to any other dual-axis combo chart (a bar-and-line combo, say): each layer keeps its own Marks card, aggregation, and Compute Using if it has table calcs, synchronized only by matching physical position on the shared latitude/longitude axis.
6. This also explains Section 5's custom-territory limitation: without a recognized geographic role, there's no key to join against the internal geocoding table at all — a custom import (Map menu → Geocoding → Import Custom Geocoding) works by adding new rows to that lookup table so the same join mechanism can succeed for territory names Tableau doesn't ship with by default.

Cheat sheet

Task	How
Confirm/set a geographic role	Right-click field → Geographic Role
Fix unmatched location values	Click the unmatched-values indicator on the field
Quick map from a geo field	Double-click the field in the Data pane
Switch dot map to filled/shaded	Marks card dropdown → Map → Filled Map
Layer points over a filled map	Duplicate Latitude → Dual Axis
Map a custom territory	Map menu → Geocoding → Import Custom Geocoding

Exercise

Using the extended Orders table with State (Section 1), build the filled map of Sales by state described in Section 3, then add a second, circle layer sized by SUM(Quantity) using the dual-axis technique from Section 4. By hand, compute total Quantity per state from the original Orders table (Module 1) matched to each state above, and confirm which state's circle should render largest.