Visualisation tools and techniques help to draw insights as well as find trends and patterns in AI/ML projects. Open source tools are convenient and low cost solutions for converting raw AI/ML data into powerful visualisations. This article lists a few of these tools.
Data visualisations play a vital role in AI/ML projects, and help to get insights of data before and after the application of any machine learning model. Visualisation allows trends and patterns to be seen more effectively and easily, and is used for numerous purposes such as to analyse changeover time patterns, to determine the frequency of certain attributes, determine correlations among data, examine clusters of data, analyse the value and risks of underlining data, and schedule projects. There is a variety of charts that can be used in visualisation for different purposes. Line, Area, Bar, Histogram, Scatter, Bubble, Pie, Gauge, Map, Heat, Frame are popular visualisation charts in AI/ ML projects.
The best data visualisation tools include both OSV (open source visualisation) tools as well as proprietary licensed solutions. Open source visualisation tools are vital as they are free of cost and provide similar kinds of functionalities as proprietary software. Table 1 describes the advantages of open source visualisation tools based on a number of parameters.
| Parameters | Benefit |
| Features/ Functionalities |
Most OSV tools consist of numerous features/ functionalities and are applicable for all aspects of visualisation |
| Complex data handling |
These tools handle complex data efficiently and are suitable for current needs of Big Data |
| API support | Most tools provide excellent API support |
| Ease of operation | OSV tools are command line or GUI based, and easy to operate |
| Development and updates | The best part about open source visualisation tools is that the development community is always making them better |
| Integration | Can be easily integrated with third party software and reporting dashboards |
| Pricing | Free of cost |
| Support | Lot of online tutorials and websites are available to ask questions from the community |
Table 1: Benefits of open source visualisation tools
OSV tools for AI/ML
- Hiplot: This is a very powerful and lightweight tool for high dimensional data. It exists as several models – as a Web server in Jupyter Notebook and as a command line interface (CLI). It represents information clearly by using features of parallel plots and other graphical utilities. It is vital for machine learning researchers as it evaluates the influence of their hyper parameters easily and effectively. This tool has various advantages in comparison with other tools. The main advantages are:
- Simple to use: It has three main variants: Web server, Jupyter Notebook and CLI, and is therefore suitable for many users. A simple syntax makes it easy to use for beginners.
- Extendibility: It provides default as well as custom parser facility. By default, the Web server of Hiplot parses CSV and JSON files. It is compatible with a number of AI libraries.
- Interactive visualisation: Due to its interactive visualisation feature, we can change the visualisation for different use cases.
- R: R is a powerful computer language for statistical computations and enriched graphical capabilities. It offers the typical range of standard statistical plots, including bar plots, histograms, pie charts, scatter plots, box plots, etc. The ggplot2() package of R is a very powerful visualisation package that uses the Grammar of Graphics concept. The latter is a framework to construct visualisation in a standard structured manner. The two major functions of visualisation in ggplot2() are qplot() and ggplot(). qplot() is similar to a standard plot() function for quick plot generation, while ggplot() provides granular control of everything in all categories of the graph. The other important package is plot3D, which contains several 3D functions. These are listed in Table 2.
| 3D function | Purpose |
| box3D | Draws the boxes between pairs of points. |
| image3D | Adds an image in a 3D image plot. |
| spheresurf3D | Plots a coloured image on a sphere |
| trans3D | Projects 3D elements to 2D. |
| hist3D | Generates a 3D histogram. |
| persp3D | Extends the 3D perspective plots defined by the persp() function. |
| scatter3D | Generates the scatter 3D plot. |
| surf3D | Plots the surface in 3D with colours. |
| rect3D | Draws a rectangle in 3D. |
| ribbon3D | Draws a perspective plot as ribbon. |
| points3D | Draws the coloured point in 3D. |
| lines3D | Draws the coloured lines in 3D. |
| isosurf3D | Plots isosurfaces from a 3D data set. |
| voxel3D | Plots isosurfaces as points. |
Table 2: 3D based functions of the plot3D package
- Orange: This is a great visualisation toolbox for novices as well as experts, and is very simple to use. No prior programming experience is required. Orange also comes with a visual programming environment and its workbench consists of tools for importing data, dragging and dropping widgets, and links to connect different widgets for completing the workflow. Orange uses common Python open source libraries for scientific computing, such as Numpy, Scipy, and Scikit-learn, while its graphical user interface operates within the cross-platform Qt framework.
- Facets: Facets is an open source visualisation tool that can help you understand and analyse AI/ML data sets. It comprises two visualisations for understanding and analysing these: Facets Overview and Facets Dive. Facets Overview automatically gives users a quick understanding of the distribution of values across the features of their data sets. Multiple data sets, such as a training set and a test set, can be compared on the same visualisation. Facets Dive provides an easy-to-customise, intuitive interface for exploring the relationship between the data points across the different features of a data set. With Facets Dive, you control the position, colour and visual representation of each data point based on its feature values. If the data points have images associated with them, these images can be used as visual representations.
- TensorWatch: This is a splendid tool to visualise data streams. The main functionality of the tool is to monitor the process of learning models in real-time in Jupyter Notebook. TensorWatch allows the user to customise a part of the model, as well as customise how the user wants to visualise and create dashboards. TensorWatch supports different types of visualisations: direct graph, pie chart, histogram, scatterplot, and 3D versions of many graphs.
To conclude, data visualisations play a vital role in AI/ML projects. The best data visualisation tools include both OSV (open source visualisation) tools as well as proprietary licensed solutions. This article describes benefits of open source visualisation tools based on parameters like features, complex data handling, API support, ease of operations, development and update facility, etc. All the tools described in this article provide great functionalities for visualisation from the artificial intelligence and machine learning perspectives.
