An (incomplete) checklist for making geodata visualizations in (data-driven) journalism.
ALWAYS ASK YOURSELF FIRST:
Is a map really the best way to visualize the data set?
- Got all geographic elements right? (especially borders & place names)
- Check the correct position of geocoded and self drawn map elements
(thus preventing mistakes from misused spatial reference systems)
- Have all outliers and duplicates been eliminated? Correctly dealt with incomplete data entries?
- Have data entries that are not necessary for the final visualization been removed?
- Have the values been normalized (e.g. by population data)?
- Has a suitable spatial reference system been chosen?
(consider rotation, distortion, equal area)
- Is the level of given context information (reference points, place names, borders) well balanced?
- Is it really necessary to use a basemap or can users already orient themselves by the mapped data points?
- Is there too much useless white space on the map? If yes, how about several smaller maps?
- Do readers instantly understand the key message of the visualization?
- Is the visualization readable for someone who suffers from color blindness?
- Choose meaningful categorial boundaries that lead to well perceivable categories
- If mapping locations: pick symbols than are more meaningful than pins (or use dots instead)
- Color scales
- Rainbow colors and red-green color schemes usually should not be the first choice. http://colorbrewer2.org/ is your friend!
- Be careful with manipulating color schemes when working with continuous data. You might damage the logic of equidistant colors.
- When working with qualitative color schemes: do all colors appear equally important or are there unintended highlights?
- Be aware of biased colors (e.g. red and green) and what meaning is usually assigned to them
- Choose a suitable middle point when working with diverging color schemes
LEGEND AND ATTRIBUTION
- Include copyright and attribution for all external data sources (incl. geodata and open data!)
- Does the reader need a legend to understand the visualization? If yes, is your legend complete?
- If you use transparent map elements, pay attention on displaying them correctly in the map legend.
- If desired: how about publishing your raw data?
- Is it easily understandable how interactive elements work?
- Would a static map be equally suitable? Does interactivity really lead to a surplus that’s worth the external dependencies that come along with interactivity?
- Is your map limited to meaningful extends and zoom levels?
- How does your visualization work with your content management systems?
- Is everything working fine in most common browsers and on mobile devices (at least in iOS + Android)?
- If desired: how about embedding options?