Visualization: Difference between revisions
Jump to navigation
Jump to search
mNo edit summary |
mNo edit summary |
||
| Line 55: | Line 55: | ||
==Misc== |
==Misc== |
||
[http://www.bogotobogo.com/python/python_matplotlib.php This is interesting] |
* [http://www.bogotobogo.com/python/python_matplotlib.php This is interesting] |
||
* See also VTK, a C++ visualisation toolkit. [https://www.slideshare.net/viz4biz/pydata-nyc-2015-presentation-55007477 via] |
|||
Revision as of 04:48, 4 March 2017
- Probably in Rust.
- Reactive/functional.
- When Data is changed the graph gets updated.
- Instead of data, can plot can be defined by a function that changes at runtime. Data is really just a mapping function anyway.
- Pub/sub? Channels?
- Can I put in a fixed value without having to manually wrap it?
- Rust generics?
- What about passing things up the hierarchy? ie, graph size changes, the axis size need to change too...
- Can the number of dimensions change?
- What about animating the changes (interpolation function between old and new)? Animation itself can be a function. How fast? Linear, quadratic? 'Fade' vs more elaborate things like 'morphing' eg a scatter plots that moves the dots, or text that changes?
- Wouldn't just be for fixed data, could be interpolating between 2 functions while they are still being updated.
- Basic things like the Axis themselves can be functions.
- Axis size/position.
- Number of 'nodes' that show numbers and such. Mark a point every 10? Split up based on the length and put a point every 1/5th. Function to generate text? Function that defines points?
- 2D and 3D?
- When to extrude a 2D graph in 3D space.
- A bar chart could be flat or have thickness.
- A function could become a surface.
- When to extrude a 2D graph in 3D space.
- How to deal with data that defines the plot points, vs plot points based on the graph.
- For example a scatter plot vs pixel plot of a continuous function.
- When you zoom in on the continious pixel one, you increase the resolution.
- Zooming in on the scatter plot 'might' increase the size of the points or not.
- A continuous function could wrap a scatter plot style one, but it would be a pain
- Would have to define circle points as a formula across the data.
- Would loose metadata about the points.
- Would be very inefficient as you would have to scan across the whole plotting canvas.
- What about allowing data to be modified by dragging points on the graph around?
- Non-square graphs
- eg, Polar graphs
- Complex numbers?
- Are the things that we draw on the 'points' on a scatter plot returned by a function (ie a circle, a triangle?). They could be small 'graphs' themselves. Ie a circle.
- Different things should be graphical 'objects' or 'graph elements'. A circle when zoomed in shouldn't pixelate. But a bitmap of a circle would. Same kind of thing with a line.
- How do we get the resolution of things like lines and circles? What about in 3d?
- Line widths for things like axis? Could be a function itself. Could be animated.
- Color as a function. Could be animated. Could be for a specific position along a line. A specific point in a scatter plot.
- Plotting things could return a 'graph elements'.
- Ie a bitmap, line, circle, etc...
- Plotting an array could return 'points'.
- Points could be interperated as a bunch of circles, a line, a smooth line (a bezier curve, function fitted to points).
Misc
- This is interesting
- See also VTK, a C++ visualisation toolkit. via