Choosing The Right Projection: Balancing Accuracy And Ease Of Use

The Dilemma of Projection Choice

Selecting the optimal map projection involves navigating complex trade-offs between preserving accurate shapes, areas, distances, and directions. The earth’s surface is three-dimensional and spherical, yet maps depict the world on a two-dimensional plane. Flattening and projecting the globe introduces distortions that vary based on the projection used.

Certain projections accurately depict some spatial attributes while severely distorting others. Cartographers refer to projections prioritizing shape as “conformal”, those prioritizing area as “equal-area”, and those focusing on accurate distances as “equidistant.” However, no single projection can perfectly preserve all these properties simultaneously.

For example, the widely used Mercator projection accurately represents angles, shapes, and directions, enabling straight-line ship navigation. However, it warps larger objects near the poles into gigantic proportions, drastically misrepresenting their size. In contrast, the Mollweide or Goode Homolosine equal-area projections display countries and continents closer to accurate relative areas, but heavily distort their shapes.

Comparison of the conformal Mercator (left) and equal-area Goode Homolosine projection (right) for the southwestern United States. Notice the emphasized horizontal stretching near the pole in Mercator.

Factors Influencing Projection Selection

Choosing an appropriate projection depends largely on the intended analysis, region of interest, data formats, and software capabilities.

Identifying Intended Use Cases and Analysis Types

Carefully consider which spatial attributes require accurate depiction for answering your research question or conducting analysis. Conformal projections work well for navigation, bearing, and directionality analyses by preserving local shapes and angles. An equidistant projection suits distance measurement best but distorts areas.

Equal-area projections prove useful for quantitative comparisons between regions and spatial statistical analysis. For example, an equal-area projection facilitates appropriately comparing population densities between countries. Using an equal-area projection prevents very large countries from visually dominating maps and better represents the relative number of people residing in each area.

Considering Region Size and Location

Certain projections work better for specific geographic extents and locations. Local conformal projections like stereographic projections accurately depict shapes and directions for mapping poles and circumpolar regions. Meanwhile, projections like Albers equal area conically project regions extending predominantly east-west like North America and Europe with minimal area distortions.

For world maps, compromise projections balance distortions globally at the cost of warping all spatial aspects everywhere. Choosing between interrupted homolographic, pseudocylindrical, or sinusoidal projections requires evaluating their distortion patterns against your use case.

Accommodating Software and Data Formats

Modern GIS software packages include extensive libraries of global and regional projection options. However, accessing some projections requires importing supplemental spatial reference systems. Additionally, certain software better facilitates manipulating or customizing projection parameters for specific use cases.

When importing external data layers, investigate their native projection and transformation pipeline. Converting projections multiple times degrades quality and amplifies distortion. Take care to project data into appropriate final map projections saving conversion steps.

Recommendations for Common Use Cases

These suggestions serve as a practical starting point when selecting projections:

Mapping Countries or Small Regions

Mercator – Preserves local shapes and bearings for navigation

Robinson – Balances global distortions for visually appealing world maps

Albers Equal Area – Minimizes area distortions for regions spanning east-west orientations like North America or Europe

Preserving Shapes

Conformal projections – Mercator, stereographic, Lambert conformal conic

Maintaining Accurate Areas

Equal-area projections – Albers, Mollweide, Eckert IV, Goode Homolosine

Keeping Accurate Distances

Equidistant projections – Azimuthal Equidistant, Equidistant Conic, Plate Carrée

Example Code for Setting Projections in Python and R

Python:

import geopandas 
import matplotlib.pyplot as plt

world = geopandas.read_file(geopandas.datasets.get_path("naturalearth_lowres"))  

# Set projection to Robinson 
ax = world.plot(projection=ccrs.Robinson()) 

plt.show()

R:

library(ggplot2)
library(sf)   

world = st_read(system.file("shape/world.shp", package="maptools"))  

# Set projection to Eckert IV equal area
ggplot() + 
  geom_sf(data = world, projection = st_crs(34))

Achieving the Optimal Balance

Finding an acceptable compromise between distortion and usefulness poses a significant challenge with geographic projections. However, various strategies improve and streamline the projection selection process.

Strategies for Evaluating Distortion Thresholds

Measure levels of angular, areal, and distance distortion introduced by a given projection quantitatively to assess suitability for your use case’s accuracy requirements. Mathematically derived distortion models and heat maps help visualize distortions across the map area.

Set acceptable distortion thresholds and maximum allowable error levels for your analysis early when choosing projections. Be realistic regarding precision needs to avoid rejected projections due to artificially high accuracy standards.

Weighing Accuracy Needs Against Ease of Use

Evaluate whether spending additional effort understanding a more niche projection saves substantial improvements in meaningful accuracy versus a simpler or standardized option. Overly complex projections provide little benefit if output maps become confusing and difficult to interpret by audiences.

Many standardized projections balance usability and distortions well for broader use cases. Expending significant effort customizing projections tends to reach diminishing returns and introduces reproducibility issues when sharing analyses.

Future-proofing with Best Practices

Utilize projection metadata standards like PROJ and spatial reference systems to precisely track projection details within data workflows. Clearly documenting custom projections improves reproducibility and understanding for later analyses.

Archiving raw geospatial data in original reference systems before projecting for mapping preserves maximum accuracy for future alternative projections. As computational analysis and visualization capacities evolve, processing the source data into new optimal projections will grow continually easier.

Carefully evaluating and documenting coordinate transformations creates quality projections facilitating accurate, actionable insights from spatial data.