Scribd Logo
Upload

Welcome to Scribd!

  • 0 views

    tSNE1

    Uploaded by

    sankeerthmanmadhan2002

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd

    tSNE1

    Uploaded by

    sankeerthmanmadhan2002
    0 views20 pages

    Copyright

    © © All Rights Reserved

    Available Formats

    DOCX, PDF, TXT or read online from Scribd

    Did you find this document useful?

    Is this content inappropriate?

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    0 views20 pages

    tSNE1

    Uploaded by

    sankeerthmanmadhan2002

    Copyright:

    © All Rights Reserved
    Download as DOCX, PDF, TXT or read online from Scribd
    Download as docx, pdf, or txt
    You are on page 1of 20

    The wine quality dataset

    Visualizing data in One Dimension (1-D)

    Histograms

    Visualizing attributes as one-dimensional data


    The plots above give a good idea about the basic data
    distribution of any of the attributes.

    Visualizing data in Two Dimensions (2-D)


    One of the best ways to check out potential relationships or
    correlations amongst the different data attributes is to leverage
    a pair-wise correlation matrix and depict it as a heatmap.
    Visualizing two-dimensional data with a correlation heatmap

    The gradients in the heatmap vary based on the strength of the


    correlation and you can clearly see it is very easy to spot
    potential attributes having strong correlations amongst
    themselves.
    Another way to visualize the same is to use pair-wise scatter
    plots amongst attributes of interest.

    Visualizing two-dimensional data with pair-wise scatter plots

    Based on the above plot, you can see that scatter plots are also a
    decent way of observing potential relationships or patterns in
    two-dimensions for data attributes.
    Use stacked bars or multiple bars for the different
    attributes in a single plot.

    Visualizing two-

    dimensional discrete, categorical data in a single bar chart


    Visualizing data in Three Dimensions (3-D)

    Considering three attributes or dimensions in the data, we can


    visualize them by considering a pair-wise scatter plot and
    introducing the notion of color or hue to separate out values in a
    categorical dimension.

    Visualizing three-dimensional data with scatter plots and hue (color)


    The above plot enables you to check out correlations and
    patterns and also compare around wine groups.
    Let’s look at strategies for visualizing three continuous,
    numeric attributes. One way would be to have two dimensions
    represented as the regular length (x axis) and breadth (y-axis)
    and also take the notion of depth (z axis) for the third
    dimension.
    Visualizing three-

    dimensional numeric data by introducing the notion of depth

    Visualizing data in Four Dimensions (4-D)


    One way to visualize data in four dimensions is to
    use depth and hue as specific data dimensions in a
    conventional plot like a scatter plot.

    Visualizing data in four-dimensions leveraging scatter plots and the concept of


    hue and depth
    One strategy to make this better, is to keep a 2-D plot, but use
    hue and data point size as data dimensions. Typically this would
    be a bubble chart similar to what we visualized earlier.
    Visualizing data in four-dimensions leveraging bubble charts and the concept of
    hue and size

    Visualizing data in Five Dimensions (5-D)

    To visualize data in five dimensions, we leverage various


    plotting components. Let’s use depth, hue and size to represent
    three of the data dimensions besides regular axes representing
    the other two dimensions. Since we use the notion of size, we
    will be basically plotting a three dimensional bubble chart.

    Visualizing data in five-dimensions leveraging bubble charts and the concept of hue, depth
    and size

    Visualizing data in Six Dimensions (6-D)


    Leverage depth, hue, size and shape besides our regular two
    axes to depict all the six data dimensions
    Visualizing data in six-dimensions leveraging scatter charts and the concept of hue, depth, shape
    and size
    Can we go higher?
    It definitely becomes more and more difficult to hack our way
    around the limitations of a two-dimensional rendering device to
    visualize more data dimensions.

    Visualizing Image Data

    Images are basically multi-dimensional tensors which can be


    represented as matrices of pixel values. This really increases the
    dimensionality of even a simple small image and makes it difficult
    to work with them

    You might also like