Three ways to visualize your data

d3js.org

Everyone loves data visualization. By converting data into art, analysts and researchers are able to translate complexity into storytelling. There are countless ways to visualize data, from making simple charts in MS Excel to custom designed javascript applications. In this post we'll talk about the three core ways to visualize data, when to use each method, and some of the associated challenges with each.

In application

The most straight forward way to visualize data is in your analysis application. Many analytical tools like MS Excel, Python, R, STATA, etc. all provide access to visualization elements that range in power, flexibility and ease of use.  In general, using these types of applications are the fastest and easiest way to create visualizations, so if you are on a deadline or producing a client deliverable, this is usually the best way to go.

MS Excel

Python

Of course, there are many limitations to keeping your visualizations inside of your analytical tools. While "power users" can do some impressive stuff with analytical tools, you still must work within the framework provided by the application. Many tools are getting better at data visualization with greater options and increased flexibility, but it's still difficult (though not impossible) to provide a visualization that "wows" your customers or clients. If you are using "big data", some of the visualization tools inside of different tools (like Python or STATA) can help you, but they require additional libraries, a strong background in the application and/or some programming skills to get great visual results. Finally, if you are looking to make your chart or visualization interactive or web-based then you are generally out of luck.

When to develop visualizations "in application"

• you are on a tight deadline
• your deliverables are static and require no interactivity (web)
• you are operating with "small data"
• you have experience in your chosen tool

Business intelligence tools

There are currently dozens of business intelligence (BI) that can be used to create sophisticated visualizations. Software applications like Tableau, Spotfire and Qlikview offer a robust set of visualization templates including mapping (GIS) and charts/graphs not available natively in standard analytical applications. Most BI tools offer free trial periods and robust tutorials, so depending on your timetable, the most effective approach is typically to select a couple and kick the tires. We also recommend checking out the Gartner Magic Quadrant as one of your first steps:

A few warnings when using Business Intelligence tools and platforms. Increased user-friendliness of BI tools comes at a sacrifice to flexibility on both the front and back ends. Most BI tools require your data to be structured in a very specific way either in your spreadsheets or database files, which can take up the majority of your project time to set up. On the front end, prefabricated templates and structures can limit creativity with your visuals. While these templates cover the needs of many users, those looking for more cutting edge or original designs may be disappointed. Finally, while many BI tools advertise themselves as fit to be used in web applications or as part of customer-facing products, their lack of flexibility and limits on user interactivity make them more ideal for internal decision making.

Tableau

When to use BI for visualization

• you are using "big data" sets
• you require mapping and more sophisticated visualizations
• you have budget for third party application
• you don't have in house front-end designer(s) or developer(s) at your disposal
• you want your visualizations to be dynamic
• your focus is internal rather than external

Open source libraries

Open source charting libraries have become some of the most popular options for visualizing data on the web. While they are not as flexible as creating visuals from scratch, open source libraries provide developers the backbone to create fun, creative and interactive visualizations that can impress any beholder. These types of libraries are the optimal choice for client-facing products and anything that is web-driven. Highcharts is great for basic visuals, and can provide solutions to charting problems ranging from the simple to the complex. D3 on the other hand is more avant garde and community-driven, so it is often a breeding ground for visual creativity. At minimum it's a great place to go for inspiration (like the one below from The Guardian or this terrific one from The New York Times).

Courtesy of The Guardian, d3js.org

The primary challenge with using these types of tools is that you need some knowledge of javascript or a professional developer who does. Additionally, you need to have your data hosted someplace (Shooju works great!) to pull it into your visualizations. It's also not optimal for short timelines, as development time can range depending on the complexity of the visualization and the data behind it.

Highcharts.com

When to use open source libraries for visualization

• your data visualization is going on the web
• you have javascript knowledge or an in-house/contract developer
• you have a longer timeline
• you have a data platform or data hosting environment
• you are looking to be creative or "wow" your customers

Conclusion: do your homework!

This post is meant simply as a starting point on different ways to visualize your data. Before you begin your visualization project, we suggest you do your homework to get a full understanding of what products and services are out there to maximize the value of the information you are trying to visualize. Best of luck!