Qualities for explanatory data visualizations

The following qualities for explanatory data visualizations were described in the free, online course Data Visualization and D3.js.

  • understand the context- who will be viewing the visualization, what is the cultural context?
  • choose appropriate type of visual- should it be a map, chart, etc?
  • identify and eliminate clutter – remove non-essential details, show only informational design elements
  • direct audience attention – show the viewer what is important, guide their experience through the visualization
  • tell a story – tie your visualization to a larger narrative

A person-friendly web of information exchange


People communicate in myriad ways. We have channels of information taking various forms such as encyclopedias, dictionaries, maps, photo albums, journals, letters, postcards, and more. Often, information exists in separate realms, and is difficult to combine. This, in turn, makes it difficult to create into rich and evolving forms.


People need an information tool(kit) that is easy to use. There needs to be little separation, or contortion, from working directly with the information we wish to remix and share.


We need to reward activities such as

  • improvement of existing commonwealth resources,
  • removing barriers to access
  • providing usefulness and/or relevance.

This, while filtering and discouraging abusive behaviour.

The Challenge

Create an open source interface that abstracts the complexities of API/data management. Enable people of all skills and abilities to create graphs and composite views of information from many sources.

The Interface

People will select information types and sources, and then connect the information on a canvas to create a composite display. These composites can be bundled and shared as well as embedded within other composite displays, in modular fashion.


The system might consist of modules to work with data. Generally speaking, there are at least three primary categories of modules.


Make request(s) for information, about the condition of information, and search for relevance.


Once information is located it may need to be modified, combined, filtered, etc.


Information may be rendered in at many stages of the process and in many forms, such as documents, tables, graphs, and multimedia output.


Brett Victor‘s work on interface design, and presentations such as Learnable Programming.

GIS Definitions from Geog 85


An attribute is a quality or characteristic of an object, observation, event, etc. Attributes can be recorded as fields in a database, columns in a spreadsheet, or through other structures and data formats. Attribute data can be classified into general types that generally align with statistical data types – i.e. category, numeric, etc.

Census block group

Census blocks are the smallest division of the U.S. census system, where information is collected from 100% of the households therein. Census blocks are clustered into groups, called block groups. Block groups form together in larger census tracts. This footprints of block groups and tracts are available, in TIGER line format, from the U.S. Census website.

See: https://www.census.gov/geo/maps-data/data/tiger-line.html




COGO is short for coordinate geometry, and is a reference language describing the geometric properties of observed objects. These observations, often provided by land surveyors, can be converted to digital representations using basic geometry components such as lines or arcs. An early implementation of COGO was the Integrated Civil Engineering System (ICES) created at Massachusettes Institute of Technology.

See: http://wiki.gis.com/wiki/index.php/COGO

Digital Elevation Model (DEM)

Digital Elevation Model data describe the Earth’s surface. The data can be structured as raster or vector format. The raster format represents the entire surface as a regularly spaced grid of points, while the vector format can contain only the number of points necessary to render the surface at a desired accuracy level. Each point, or node, in the data describe the elevation of the Earth’s surface at that particular place.

Geographic coordinates

Geographic coordinates assign values to points on the Earth’s surface. The values can be combinations of up to three numbers, usually indicating latitude, longitude, and altitude. The coordinates are overlain on a surface representing the entire surface, or a subset, of Earth. Different shapes, called datums, can be used to represent the Earth’s surface. This produces variability in accuracy between points on different datums, where the same coordinates on two datums can be meters apart.


Georeferencing involves associating spatial or geographic locations (i.e. coordinates) with data or objects. This can include indicating locations of objects depicted in images, such as aerial photographs.

Map projection and datum

A map projection is a process whereby the spherical surface of the earth is unwrapped and flattened to display as a map. The projection is a mathematical process that assumes a generic shape for the surface of the Earth, such as a sphere or ellipsoid, and then transforms each point on the surface to a point on a two dimensional grid. A projection can be thought of, and is analogous to, the shadow between your hand and a wall when shining a flashlight on the side of your hand opposite to the wall. There are hypothetically unlimited ways that a surface can be projected, and each method introduces some distortion in the process – creating inaccuracies in the projected result.

Map Scale

Map scale describes a ratio of one map unit as it relates to real world units of the same measurement. E.g. how many real world inches are represented by one inch on a map.


Metadata are data about data. Metadata describes characteristics of data that are not directly indicated or inherent. This can include projection/datum information, data quality/lineage assessment, attribution, licensing, and derived characteristics such as number of geometries, spatial extents, distribution of values, etc.


When taking images, including aerial photographs, there are natural distortions that occur due to perspective and optics. With aerial photographs, distortions such as perspective/tilt, lense distortion, and topographical differences, can be corrected to produce a ‘flattened’ image. This process is called orthorectification and the resulting images can be more accurately used in conjunction with other spatial data for base layers, etc.

See: http://wiki.gis.com/wiki/index.php/Orthophoto