In Bloom, you can set the Style of a Node (color, size, icon).
It would be nice if Bloom showed a preview of the node, including the text color on the node. If the color is light, the text is black. If the color is dark, then then the text is white. But, you have to use trial and error to figure out how it's going to come out.
I found the text color being another helpful visual cue as to what sort of node I'm looking at.
Thanks for the feedback. We do have logic that determines whether the text color is white or black, and this is dependent on the node color.
Would you say that the logic not always optimal? Or that you would like a setting to customize the text color to anything you please?
Cheers
I do see there is logic that determines the text to be black or white. However, there is going to be a "valley" where neither white or black will provide enough contrast (at least for me!)
Having text with optional color (and style) would be a useful addition. This is because with a complex data set, it helps a person browsing through the data to be able to see more details immediately.
Another possible distinguishing feature to add is shape and thickness and style of Node boundary. Also, style of Relationship boundaries: dotted, dashed, etc.
Since Nodes can have multiple Labels, perhaps an option to allow a Node to be multiple colored based on the number of different Labels it has would be useful.
I'm a big fan of Books by Edward Tufte and Richard Saul Wurman. As I recall, Wurman points out there are only a few ways to organize information, Alphabetically (which RDBMS can do), color, shape, size (and maybe a few others that I can't recall), which RDBMS can't do.
The human eye is remarkably good at finding patterns, so it would be great if Bloom gave the user more possible ways to see the hidden patterns in the data, i.e. how does one find information amid all the data?
Finally, another suggestion: when the nodes are too small for visible text, you can still optionally indicate the length of the text with dots (or something.). I think this would be a good idea, as a person well versed with the data could possibly see a pattern in the length of the text. Maybe even indicate descenders and ascenders of text, so that an ascending letter (e.g. "k") would be two pixels tall and a descending letter (e.g. "p") would have a pixel below the level. You could even have a gray level for the pixel depending on its rareness, so that "q" would be dark and "e" would be light. I bet people would learn how to distinguish two Nodes in their data set that way.
It's all about getting information out of the data.
[added]. I do recognized these ideas are a level of complexity that beginners shouldn't have to deal with. So, it would be a good idea to keep this out of the Browsers and only have it in Bloom.
Thanks for all this detail!
I totally agree with you. When somebody using Bloom can maximize the extraction of information by just looking at the visualization - we will know that we have made good product design decisions.
Your suggestion about text abstraction is certainly an innovative idea. Great feedback.
On a different note, I would love to hear more feedback about your experience of using Bloom. Would you be up for a 30 minute call sometime?