A demonstration of scatterplot with force directed points, so that maximum overlap is avoided. Furter, each of the axes can be filtered dynamically from the input fields provided at the top of the plot.
Clustering random points based on their similarity (groups) as star shaped spirals. fun.
This is related to the Clustering - I
example, except that it clusters related points in the form of spirals.
Example of clustering randomly placed particles using a given key (or group in this case) in the form of squares. The basic logic is decrementing the y on every cutoff point of x, where the cutoff point is determined by the number of particles allowed in a given row.
This animation demonstrates the construction of a square spiral. The logic is to add a system of kicks: at each vertex of the 4
sided growing spiral, increment the number of nodes to be added in the following vertex.
Implementing an idea of visualizing the classification
problems (with real-valued attributes) as heatmaps. Each row, in this heatmap is an attribute, where columns correspond to the classes
to be classified.
A good benefit of this approach of multidimensional visualization is that it color-codes (or exposes) key variations of different attributes (columns) over the classes
(rows) and allows the individual to decide about which attribute should be more focused, while classifying the data.
For the sake of visuals, an implementation is done on popular IRIS dataset.
The famous Sierpiński Gasket!
Idea (in simplest words) is to create equilateral triangles within a parent equilateral triangle, in such a manner that after each iteration, three more small triangles are made that perfectly fit in the parent triangle.
In this implementation, the max number of triangles you see is: 9840. This number is generalized based on the formula [(3^n)/2] - 1
, which at each n
, corresponds to a round of iterations which fill the entire parent triangle with equally sized child triangles (n
is kept to 9 in this example).
Implementation of a Sierpiński carpet. The formal procedure of construction begins with a square, which is cut into 9 congruent subsquares in a 3-by-3 grid, and the central subsquare is removed.
However, in this implementation, the problem is approached a bit differently: Each square is divided into 9 equally sized small squares (i.e. 3x3 grid), where the centroid of this grid is colored. Then each of these small squares are repetitively divided until a finite limit. At each round of iteration, 9 new squares are added in the list to be processed accordingly.
A fractal of the form in which an equilateral triangle gets sliced into two right-angled halves, and each of these halves gets a child equilateral triangle which is processed in the similar fashion.
Fractal of a form in which an equilateral triangle's base is divided by two child equilateral triangles, and the process is repeated recursively with those children.