Choropleth maps, dasymetric maps and cartograms

Detail

What a choropleth map is

• Choropleth map (from the Greek choros = place, plethos = value): a thematic map in which each administrative unit (state, district, tehsil, country) is shaded in graded tones. The shade shows the average or ratio for that unit.
• Typical variables: population density (persons per sq km), literacy rate (%), population growth rate (%), sex ratio (females per 1,000 males).

• Boundaries are fixed in advance. The data fits into existing units; the map does not draw new ones.

• It is the most common thematic map in Census atlases, district handbooks and school textbooks.

What you need to make one

• A base map of the administrative units (for example, states and UTs of India).
• One value for each unit for the same variable and the same year (for example, literacy %, Census 2001).

Steps (NCERT method: equal intervals)

1. Arrange the data in ascending or descending order.
2. Find the range: range = maximum value − minimum value.
3. Find the class interval: range ÷ 5. This gives five classes.
4. Name the classes very low, low, medium, high and very high.
5. Shade from light (very low) to dark (very high). Use shades of one colour or one pattern, so that darker always means "more".
6. Add a legend showing each class and its shade, plus a title, scale, north arrow and source.

NCERT worked example: literacy, Census 2001 (Table 3.10, Fig. 3.14)

• Lowest: Bihar, 47.0%. Highest: Kerala, 90.9%.
• Range = 90.9 − 47.0 ≈ 44. Interval = 44 ÷ 5 ≈ 9.

Class	Range (%)	Shade
Very low	47–56	Lightest
Low	56–65	Light
Medium	65–74	Middle
High	74–83	Dark
Very high	83–92	Darkest

• Very high: Mizoram (88.8%) and Kerala (90.9%). Lakshadweep (86.7%) also falls here. NCERT's class lists leave out several UTs, so this is easy to miss.
• Very low: Bihar (47.0%) sits at the bottom of the scale.

Same method, newer data (Census 2011)

• In Census 2011, all-India literacy was 72.98%. Kerala was highest at 94% (male 96.1%, female 92.1%). Bihar was lowest at 61.8% (male 71.2%, female 51.5%) [3]. (NCERT: Bihar 47.0%, Kerala 90.9%, Census 2001.)
• Literacy rose from 18.32% in 1951 to 72.98% in 2011 [3].
• Gender gap in literacy, Census 2011: smallest in Meghalaya (3.1 percentage points), below 5 points in Kerala and Mizoram, and largest in Rajasthan (27.1 points) [3]. This variable also suits a choropleth map well.
• Applying the NCERT steps to 2011: range = 94.0 − 61.8 = 32.2, interval ≈ 6.4. The classes become roughly 61.8–68.2, 68.2–74.7, 74.7–81.1, 81.1–87.6 and 87.6–94.0.
• The ranking stays the same (Kerala top, Bihar bottom).
• The class limits move because the whole country moved up. A "medium" in 2011 is not the same number as a "medium" in 2001. Always read the legend.

The golden rule: ratios, not absolute counts

• Use ratios (per cent, per sq km, per 1,000) and never raw totals.
• Why:
• Raw totals grow with the size of the unit.
• Large states such as Uttar Pradesh or Rajasthan would look "high" just because they are big or have many people.

• The map would then show size, not the real intensity of the variable.

• Absolute counts belong on dot maps or proportional symbol maps (circles or spheres), not on choropleths.

Limits of choropleth maps (beyond NCERT)

• Uniform value inside each unit:
• The whole unit gets one shade, so the map suggests the value is the same everywhere inside it.
• In reality a district has dense towns and empty forests.

• Cartographers call this the assumption that the phenomenon is evenly spread across the enumeration unit (enumeration unit: the area for which data is collected, such as a census district) [2].

• Abrupt change at boundaries:

• Values seem to jump suddenly at a state or district line.

• Real change is usually gradual. This is called the artificial transition at the boundaries of mapping units [2].

• Large, thinly peopled units dominate the eye:

• Ladakh, Rajasthan and Arunachal Pradesh (India), or Russia, Canada and Australia (world), cover huge areas but hold few people.
• Their colour fills most of the map, so the reader over-weights them.

• Small but crowded units such as Delhi, Chandigarh and Lakshadweep almost vanish.

• Class choice changes the story:

Classification method	How classes are made	Effect
Equal interval (NCERT method)	Range ÷ number of classes	Easy to read. Many units may fall in one class if the data is bunched.
Quantiles	Same number of units in each class	Every shade is used. Very different values may share a class.
Natural breaks (Jenks)	Breaks placed at natural gaps in the data	Fits the data well. Classes differ from map to map, so maps are hard to compare.

• The same data can look "mostly low" or "mostly high" depending only on the method chosen.

• Simplified binary shading can mislead: an election map may shade a whole state in one party's colour when that party won only 50.1% of the vote. This ignores the other 49.9% of voters [5].

Remedy 1: dasymetric map

• Dasymetric map (Greek dasys = dense, metron = measure): a map that redraws the zones using extra information, called ancillary data (extra data that helps place the values), so that values sit where the phenomenon actually exists.
• How it works (cause → process → result):
• Cause: a census gives one population figure for a whole district.
• Process: overlay land use and land cover (LULC) data. Remove uninhabited parts (forests, deserts, water bodies, glaciers). Put the population back only on settled and farmed land [2].

• Result: new, more uniform zones that show real density changes inside the unit [2].

• The USGS released GIS software in 2007 to rework census data with dasymetric methods based on LULC [2].

• Where it helps in India:
• Districts with large empty areas: the Thar Desert (western Rajasthan), the forested Western Ghats and north-eastern hills, the Rann of Kutch (Gujarat), and the Sundarbans mangroves (West Bengal).

• In these places a plain choropleth spreads people over land where nobody lives.

• Why use it: to show population distribution more accurately, which matters for disaster exposure, service planning and resource allocation [2].

Remedy 2: cartogram

• Cartogram (also called a value-by-area map): a map in which each unit's size is drawn in proportion to a variable (population, GDP, voters, emissions) instead of its land area [6].
• A large value swells the unit and a small value shrinks it [6].

• Shapes, directions and areas are bent on purpose to make the statistic easy to see [6].

• World population cartogram:

• China and India swell, because each has more than one billion people.
• Russia and Australia shrink to thin strips [4]. Canada shrinks too.

• Africa and South Asia look much larger than on an ordinary map.

• Other uses: a cartogram of arms exports would show a swollen United States (the world's top arms exporter) [6]. GDP cartograms swell the USA, China, Western Europe and Japan.

• Election maps:
• Constituencies are drawn the same size (each seat = one block, a hexagon/tile cartogram), or sized by number of voters.
• Crowded urban seats (Mumbai, Delhi, Kolkata) then get fair visual weight against huge rural seats such as Barmer (Rajasthan) or Ladakh.

Choropleth vs dasymetric vs cartogram

Feature	Choropleth	Dasymetric	Cartogram
Zones used	Fixed administrative units	New zones redrawn using ancillary data (LULC)	Administrative units, resized
What varies	Shade or tone	Shade inside refined zones	Size (and often shade)
Geography	True shape and area	True shape and area	Shape and area distorted on purpose
Data	Ratios or averages	Ratios or density, redistributed	Absolute totals (population, GDP, votes)
Main weakness	Uniformity, boundary jumps, big empty units	Needs good ancillary data	Hard to recognise places
Typical use	Census literacy or sex-ratio maps	Population density, disaster exposure	World population or GDP, election results

Prelims Hooks

• A choropleth map shows ratios or averages (density, literacy %, sex ratio) for administrative units, not absolute counts. Absolute counts go on dot or proportional-symbol maps.
• NCERT steps: arrange data → range = max − min → interval = range ÷ 5 → five classes → shade light (very low) to dark (very high).
• NCERT literacy map (Census 2001): Bihar 47.0% (lowest) to Kerala 90.9% (highest). Range 44, interval ≈9. "Very high" (83–92): Kerala, Mizoram (88.8%), Lakshadweep (86.7%).
• Census 2011 literacy: India 72.98%, Kerala 94% (highest), Bihar 61.8% (lowest). Smallest gender gap is Meghalaya and largest is Rajasthan [3].
• Dasymetric map = zones redrawn using ancillary data (land use/land cover). Forests, deserts and water are removed before density is mapped [2].
• Cartogram = value-by-area map. Size is drawn in proportion to a variable, not to land area [6].
• On a population cartogram, India and China swell and Russia, Canada and Australia shrink [4].
• Trap: "A choropleth map shows how values change inside an administrative unit." False. It assumes one uniform value per unit [2].
• Trap: equal interval, quantile and natural-breaks classes made from the same data can give different-looking maps.

Mains Points

• Map choice shapes policy:
• A state-level literacy choropleth (Kerala 94% vs Bihar 61.8%, Census 2011) [3] hides gaps inside each state, such as female literacy (Bihar 51.5%) or rural–urban gaps.

• District-level or dasymetric maps are needed to target schemes at the right blocks.

• Disaster and resource planning (GS-III):

• Dasymetric maps built on LULC data place people where they actually live [2].

• This gives a better estimate of who is exposed to floods (Brahmaputra and Kosi plains), cyclones (Odisha and Andhra coasts) or heatwaves than whole-district averages do.

• Visual bias in world maps:

• Area-based maps over-weight thinly peopled giants (Russia, Canada, Australia).

• Population cartograms show the true human weight of South Asia and China [4]. This matters when discussing climate justice, per-capita emissions and global inequality.

• Honest data communication:

• Winner-takes-all election shading hides close margins (50.1% vs 49.9%) [5].
• Cartograms and finer classes give a fairer picture. This is a good point on the ethics of thematic mapping in GS-I human geography.