By Dan Lamyman
Co-Founder & Director
Business Intelligence & Advanced Analytics
Data visualisation is not, in itself, a new thing. Charts and graphs have been in use throughout the ages, and you’ve been able to pull them together in Excel for a number of years now. What’s changed in recent years is the amount of data we have access to, the kinds of data available, and the tools to process it, to visualise it, and to share it.
As a result of this new profusion of data, it’s little surprise that a wealth of data visualisation tools are now available. However, each has its own specialist areas, things that one is better at than another and vice versa. The key is finding the tool that offers the features you require for your own purposes. As such, a rundown of the best on the scene and the ways they can be used seems to be a good way to get started.
Let’s begin with the so-called ‘Grand Master’ of data visualisation tools. Tableau boasts over 57,000 user accounts and is used across many industries. The interactive interface far surpasses the visualisations that can be achieved with general business intelligence solutions, whilst the simple drag-and-drop functionality speeds up and simplifies the creation process.
Tableau can easily handle large, fast-changing datasets, including machine learning applications (a feature that will ensure the platform stands the test of time as ML becomes an increasingly central part of business intelligence).
The platform integrates smoothly with Hadoop, Amazon AWS, MySQL, SAP, and Teradata.
Your account offers you 10Gb of space on a customisable public profile, with both free and paid versions of the software available. The free version is superb and widely used, however it comes with a large, Tableau-branded footer which can only be removed by migrating to the paid version. Both are easily embeddable into websites and blogs, remaining live and interactive in situ. Equally, sharing to social media is a simple process from Tableau or direct from your website.
Use when: You have AI and/or fast-changing datasets, want to share your data visualisation on social media, or to showcase your work on your website.
Languages: C, C++, Java and Python. It’s important to note that Tableau SDK only supports Python 2 and doesn’t work with Python 3.
What makes it different: Tableau’s focus on external sharing and embedding makes it perfect for outward-facing presentation more than for internal use.
Best for: Experienced users
Use when: You have the time and resources for training and for when analytics reporting is important to you.
Languages: VBscripting, SQL Server & C++, are few of languages. Excel (VBA) Object Orient Programming language and SQL knowledge will be very useful.
What makes it different: There are fewer limitations with Qlikview compared to some competitor data visualisation platforms, mainly because of its no-query approach to analytics. Instead, it uses its own Qlik Associated engine to pull data together.
Best for: Experienced users or where training is given
FusionCharts is also lauded for its vast array of visualisation choices. It includes 965 maps, 90 different chart types, and over 800 live example templates (hosted in JSFiddle). These live example templates eliminate the need to create charts from scratch, as you can plug in your own data to feed directly into the template.
The platform comes with open-source plugins for libraries like jQuery, and frameworks such as AngularJS and React. It supports JSON and XML data formats, and exports visualisations in .png, .jpg, .svg and .pdf with one click.
Use when: You need a watermark-free free platform for non-commercial work (an enterprise licence, incidentally, is $1,000 a year). A handy platform for those still working off old browsers.
What makes it different: Unlike many other platforms, all FusionCharts charts have scrolling capability, and there’s the option for both cylinder and thermometer charts which aren’t always available elsewhere. There are also node diagrams for network simulation applications and organisation charts.
Best for: Experienced users
Highcharts is used by more than 80% of the world’s largest companies. On GitHub, you’ll find a whopping 2.6 million Highcharts code references. This success may be at least partially due to its ease of use. It is one platform that requires minimal training in order to get the hang of.
Its accessibility doesn’t end there, offering perhaps the most sophisticated support for visually-impaired users and for those with keyboard-only navigation. This accessibility support means Highcharts exceeds in both Section 508 requirements and WCAG 2 guidelines.
Though Highcharts requires a licence for commercial use, which isn’t uncommon, its free trial is perfectly sufficient for non-commercial users, such as non-profits, bloggers and for school sites.
Don’t be fooled, however, by its accessibility and ease of use. Highcharts is also powerfully extendable and pluggable for experts who require advanced animations and functionality. For those creating stock charts, there’s a sister package, Highstock.
Another great thing about this platform is its touch-optimised charts, which make it ideal for mobile and tablet experiences, including touch-drag for data inspection and multi-touch for zooming. Its cross-browser support adds to Highcharts’ ease-of-use value, and its ability to automatically find optimal placement for non-graph elements, such as legends and headings, does much the same.
Use when: You have a mixture of non-techie users and a team of experts (to build deeper use).
Languages: .Net, PHP, Python, R, and Java, as well as iOS
What makes it different: Its vast accessibility support
Best for: All users
Datawrapper is a younger platform that is most certainly on the up. It has proved itself a favourite with media organisations, being built specifically for journalists, who love that it allows the embedding of live charts directly into their articles. Datawrapper is an online tool (great if you are a bit averse to downloading software to your device) with a simple, clear interface that’s simple to use without any coding knowledge or design skills. It uploads .csv data easily and swiftly, as well as allowing you to simply copy-paste data into fields from Google Sheets or Excel. However, it’s worth noting that there is very little support for editing the data, meaning the data preparation must be done yourself, which can be a nuisance.
Use when: You work in media and already have internal systems to handle data preparation. Also, if you simply need quick visualisations without any more complex control.
Languages: No coding required.
What makes it different: Its ease of use for entry-level staff, as well as its focus on media applications.
Best for: Beginners and non-techies
Plotly is a web-based data visualisation tool that has a sister software, Plotly On-Premises, for desktop use. Any chart made on another platform, such as matplotlib or ggplot2, can be made interactive on Plotly.
Its user-friendly, sleek-looking, professional-feeling interface, customisable axes, notes, legends and layout, and built-in social sharing capability are all factors that add to Plotly’s popularity as a data visualisation platform.
Use when: You require an SQL platform that feeds data in easily with your visualisation tool, i.e. when you use the platform heavily.
Languages: Python, R, MATLAB
What makes it different: Its flexibility for use with different skill levels, and for varying requirements.
Best for: A staff of mixed abilities.
Sisense is a meaty, full-stack data analytics platform that gathers data from multiple sources into one place. It supports large datasets and offers real-time dashboard queries for speed and simplicity. Sisense’s drag-and-drop interface allows for the creation of everything from simple charts through to complex graphics visualisations, all interactive, which can be shared across organisations. It’s already well-versed in AI and machine learning analysis, and integration with IoT is all there, setting the platform up for a bright future.
Use when: You have large datasets that require a full analysis platform
Languages: Custom SQL
What makes it different: Its built-in IoT, AI, and ML capabilities.
Best for: Both beginner and advanced users.
So, these are my pick of the pack for some of the best platforms out there for data visualisation. I haven’t even got on to Google Data Studio, Microsoft Power BI, or IBM Watson Analytics, though these big names do beg a similar comparison article in themselves. With the data landscape quickly changing to incorporate more AI and machine learning, it’s clear that platforms that offer advanced analytics imbued with AI capabilities will be the most successful. Sisense and Tableau are already, as we have seen, ahead of the game. IBM Watson Analytics is also sure to be a key player here.
Along with AI, the need for interactive maps is on the rise and will, no doubt, continue to do so. Why? Well, location data is becoming ever-more deep and available, so platforms that aren’t already equipped with map templates will need to up their game. The rapid increase in data, in general, is also hastening the demand for data exchanges, marketplaces, and open-source data, opening the doors for more areas for platform integration.
Data visualisation is set to move beyond simple storytelling, with increasing demand for a more interactive experience for the communication of complex issues. As such, that interactive element will be indispensable, particularly in outward-facing visualisations that are designed to engage the public. Socially-shareable visualisations that engage audiences with demonstrable data to counter the ‘post-truth’ epidemic are critical right now, particularly for journalistic and media use.
We’ve looked at some platforms that are putting accessibility front and centre of their offering. This accessibility will be in stronger demand as time goes on, as familiarity with data becomes an integral part of more job roles. IBM predicts, according to Carto.com, a 39% rise in demand for data engineers and scientists by 2020. Meeting that demand may be difficult, as the numbers of specialists coming through higher education to the workplace are currently insufficient, with the best often being snapped up by leading corporations before they can set foot in an interview. As such, more of us will need to be au fait with data as part of our everyday work, pushing the need for more platforms that make data analytics and visualisation accessible for all to use.