How to use a specialized display to optimally highlight signal, not noise?
NOAA's "Science on a Sphere" - at Berkeley's Lawrence Hall of Science, Seattle's Science Museum, and many others - is a planet-shaped screen for displaying planetary "spatial" data for earth and other planets, as well as, presumably, displaying anything that a flat screen can show. Here's the FAQ.
The SoS at the Lawrence Hall of Science offered visitors a plethora of 30+ datasets (there are more) to view; but for climate change, only one, under "models and simulations"*, offered a clear view of existing historical (not model, not simulation) year-by-year temperature data from the mid-1800s to present; then it simulated further increased temperature to 2100.
*Would a climate doubter choose to view the historic data, if they think it's just "models or simulations"?
While there was a dataset visualization showing Arctic sea ice extent, it wasn't one that would reveal the historic declines we've been experiencing, it just showed day-by-day changes over one year.
It's worth considering how NOAA could create more user-informative visualizations from its datasets. Showing Arctic sea ice minimum extent from 1979-present, then pausing to let the viewer reflect (rather than restarting the loop immediately, as I think was done with the temperature dataset) would be helpful.
Maybe the path forward would be to interview those whose data it is, & ask "what are the most important take-home message(s) from your data?" - and then design the visualization - and add informative text or audio to explain it - so it allows the viewer to see this.
My impression, from watching: When you're just shown the raw spatial data, particularly if it's labeled "A1B1" or something equally opaque to the museum visitor, it's not meaningful; beyond the labeling, perhaps Tufte or the Frankel&DePace's “Visual Strategies” book (or website) might be worth consulting, for how to display the data in a way that's more meaningful to visitors.
How can you know if you're succeeding? How about having an exit quiz, offering a token that's good for a dollar off the next visit? (LHS has an office of educational assessment, so might be a great partner for such a project.) If you don't assess the outcome, outreach can derail.
(Another idea - you could show a series of related datasets, & educate viewers about how someone wanting to minimize climate change can hide a real shift by picking a metric to display that obscures it, or someone can highlight the shift by picking one that emphasizes it. For example, show [the existing SoS dataset of] arctic sea ice day-by-day over a year, then show arctic (annual+multiyear) sea ice extent minima 1979-present, then show. arctic multiyear ice (extent?) minima 1979-present.)