Last November, we discussed our favorite data visualization tools that can be used without any licensing fees (both open source and public software, see “5 of Our Favorite Free Visualization Tools” at fulcrumanalytics.com/blog-posts/). As we begin to close in on the last quarter of 2018, we wanted to supplement what we covered 10 months ago with additional emerging tools.
Our Favorite Free Visualization Tools Vol. 2
September 20, 2018
Last November, we discussed our favorite data visualization tools that can be used without any licensing fees (both open source and public software, see “5 of Our Favorite Free Visualization Tools” ). As we begin to close in on the last quarter of 2018, we wanted to supplement what we covered 10 months ago with additional emerging tools: Dash, Superset and Vega.
Made public last year, Dash is essentially a framework that simplifies the creation of dashboards and web applications. It is an alternative visualization deployment framework based in Python that is comparable to R Shiny. With the trend of modeling and integration work migrating to Python, there is growing interest in Dash. With gained proficiency in Dash, a data scientist can perform the full spectrum of analytic activities (e.g., backend development, statistical computation, automation, and visualization) in one language. This is a key visualization tool to begin taking advantage of for 2019.
Image source: Great Ball of Fire by Ivan Nieto (https://great-balls-of-fire.herokuapp.com)
Originally developed and put into Apache as an open source tool, Superset is a great option for someone who is looking for a less complex way to create an interactive solution. Just like any other pre-built application, Superset makes it easy to explore, experiment, and generate visualizations. The application is built in Python, but simply requires the end users to write SQL queries to create dashboards. However, you may find it difficult to accommodate customized visual needs.
Image source: Superset: Airbnb’s data exploration platform, Medium (https://medium.com/airbnb-engineering/caravel-airbnb-s-data-exploration-platform-15a72aa610e5)
Developed by the Interactive Data Lab (IDL) at the University of Washington, Vega is a D3-powered* framework which allows for end users to create sophisticated data visualization outputs. It intakes JSON files and produces D3 graphics without any coding requirements. Though it is an up and coming framework, it is popular with an active community of users in Google and Slack that work to solve problems and make improvements. If you are continually seeking cutting edge visualization tools, Vega is definitely worth exploring.
Image source: Visualizing the Feed Scene on “The Ave” (p121), Alaska Beyond Magazine (http://www.paradigmcg.com/digitaleditions/abm-0917/html5/)
Image source: Beauty in Simplicity - Visualizing Large Scale Genomic Data, Aridhia (https://www.aridhia.com/blog/beauty-in-simplicity-visualising-large-scale-genomic-data/)
With the ability to display graphs, charts, maps and more, Tableau Public is a popular data visualization tool that's also totally free. With up to 10 GB of storage and a drag-and-drop interface, users can watch their data update in real-time while collaborating with others on their team. The “public” portion of Tableau means that you can only save your data to public profile where others have access to your data, but if you’re not a highly public company whose privacy is your #1 concern, there are a plethora of upsides to Tableau Public for business analysts and managers. The newest version is optimized for mobile devices, can connect to a variety of data sources beyond Excel, and can link directly to Google Sheets.
Image source: Make a Difference with Data (http://makeadifferencewithdata.com/vizs/)
Datawrapper is a great open source tool for the complete visualization of data and the ability to embed live and interactive charts. Simply upload your data in a CSV file and the online tool is able to build customized visuals such as bar and line graphs. Datawrapper is great for small business or presentation use, as it allows for only 10,000 views per chart, but it may not be ideal for big businesses with a large clientele. However, most people agree that the easy to use interface and ability to quickly present statistics in a straightforward manner is helpful.
Image source: Be a Patriot, Eat Less Beef, Mother Jones (https://www.motherjones.com/environment/2014/07/american-meat-consumption-changing-better/)
Pivot is an intuitive UI designed to enable exploratory analytics on event data while utilizing the much appreciated drag-and-drop interface. One of the attributes that sets Pivot apart is that it’s centered around two operations: Filter and Split. Filter narrows the view of data and is equivalent to the “WHERE” clause in SQL, where as Split is very similar to SQL’S “GROUP BY” function. However, Split allows for data to be cut across multiple dimensions -- we’ve seen great success in grocery price/promotional analysis and optimization.
Image source: Make Dynamic Dashboards using Pivot Tables & Slicers, Chandoo.org (https://chandoo.org/wp/dynamic-dashboard-video-tutorial/)
Data visualization and beyond...
Across a variety of industries, data visualization and interactive tools are becoming essential in daily operations. When deciding between an open source tool or a SaaS solution, we recommend first considering long term analytic goals. Data science doesn’t exist in a vacuum, and data visualization is one of the many data science outcomes. As such, the following key points are crucial in determining the best path forward:
- End user needs and their user stories
- Required complexity of the desired solution
- Current in-house data science capabilities / maturity
- Future organizational strategy for data science capability building
Regardless of an organization’s data readiness, making the most of data through visualization is a great way to showcase the benefits of applied data science. Fulcrum’s DSAT is a highly efficient and low commitment business relationship that can help companies assess and reach data science maturity. Contact us here for any questions, comments or inquiries.