Data visualizations have never been more powerful. After Tableau Conference 2021, we feel inspired by new possibilities and enhancements that will enable the world to use data more responsibly and efficiently. Here are trends we’re seeing and reasons to get excited about the future of data viz.
1. Diversity, inclusion, and ethics are at the forefront
These important topics were put at center stage during TC21. We heard from a diverse #datafam community and saw diversity in Tableau speakers and an increase in the voices of women. We also saw an increase and prevalence of important topics like data’s impact on homelessness and police budgets, and ethics in AI. While there were still plenty of tips, tricks, and tools related to Tableau, the shift to a more holistic and diverse offering of sessions was welcomed.
Data is inherently human: Ethics in AI and analytics
When we talk about data, whether intentionally or not, we sometimes leave out certain groups and also improperly group various populations. For example, voice AI at the beginning was very poor at understanding non-standard English speakers, possibly due to its training pool being too small and all native English speakers. Who decides what the training pool should look like? How do we go about ensuring the AI we develop is responsible, inclusive, and well designed?
As a starting place, it’s critical to follow these principles of ethical AI:
- AI must be responsible
- Al must protect human rights
- AI must protect the privacy of users’ data that companies collect
- AI must be inclusive of everyone impacted by it
Data is not neutral—it’s collected, analyzed, and visualized with implicit bias. This means you shouldn’t let the data speak for itself. Designers have a responsibility to treat data with respect and care, and show the data with empathy. We must be intentional and mindful of who we’re reaching out to and seeking feedback from. Is a certain feedback forum only used by certain types of people? Does your data do no harm?
All AI should be ethical AI. The more accurate and less biased the data is, the more ethical the AI solution will be, which is why leaders should incentivize designers to be as accurate as possible. Data viz is a powerful tool. Visualizing data that’s biased to groups of people can—in a very real way—have a negative impact on people and the world.
Tableau Foundation: Democratizing police budgets
In Washington, DC, where policing disproportionately impacts Black residents who are overrepresented at every stage of the legal system, the DC Fiscal Policy Institute created a budget data visualization tool. The budget tool provides a look into the police department’s funding to help DC advocates, lawmakers, and residents better understand where the funding comes from, where it goes, and why it matters.
Stakeholder feedback throughout the process in such a public-facing and sensitive topic was critically important. This includes not only feedback from the public sector and police department—ensuring the data was accurately displayed—but also stakeholder feedback from the community and target audience, ensuring the visualizations are easy to use and understand. To build additional transparency and confidence in the process, the DC Fiscal Policy Institute hosted “train the trainer” sessions around the data-gathering procedures for leaders and groups that were interested.
Lessons learned:
- Partnership is pivotal. The partnership between public officials, public sector agencies, local community leaders, and not-for-profits required a lot of thoughtful work and resources but was instrumental to the project’s success.
- Be prepared to adapt to data limitations. In this case, data limitations included receiving messy data from Freedom of Information Act (FOIA) records and having to clean publicly available data to match up with the budget data. Create a plan for how to tackle data challenges.
- Explain—don’t assume. It was important not to copy and paste an official agency document into a dashboard assuming the audience understood different budget terms. When building a dashboard, add explanations for abbreviations and acronyms.
For more details on publicly gathered or publicly available data, Tableau Public is a great resource to leverage.
2. New enhancements provide more efficiency and effectiveness
There are multiple entry points for analysts and IT experts to leverage tools in the Tableau ecosystem to increase efficiency in dashboards, prep flows and scheduling, and server/online deployments. Recent updates to Tableau Desktop, Prep, and Server provide increased speed and reliability. Here are some key product updates to be aware of and what they mean for users.
- Visualization extensions: No more late-night googling and reverse engineering Tableau Public vizzes or security concerns from IT about needing to install extensions. With these exciting new extensions to use with Tableau’s out-of-the-box options, new and seasoned developers can quickly take a unique visual and adapt it to their data.
- Dynamic layouts: Customers can now hide visualizations when there isn’t enough data to populate them. It allows developers to use a lot more space in a dashboard.
- Spatial data integration and analysis: Tableau is expanding these features—including offering advanced geospatial calculations and additional mapping tools—to compete with GIS software.
- Workbook optimizer: When you want to publish a workbook, there’s an option to run the workbook optimizer. Tableau will scan the workbook and show users potential areas of improvements, like unused data sources, a large number of dashboards (or dashboard objects), or long-running calculations that can be improved. This automated tool evaluates areas of improvement for speed, data source size, calculations, and more.
- Tableau-Slack integration: This is the year of the Tableau and Slack integration. Dashboard sharing and prep updates in Slack are just the beginning of potential opportunities between the two separate tools.
Salesforce has gone full-throttle with integrating Slack into everything—from taking full advantage of users that are in a “work from anywhere” environment to quickly pinging users to notify them of an update that they’d usually have to access via email or through Tableau online/server. - Data Change Radar: This feature monitors dashboards behind the scenes to identify outliers or anomalies. Users are notified of unexpected data changes and can easily view what those data changes were, visualize the change in the side panel, and use data to explain the change.
- Explain the viz: Opening up a new visual and understanding it has never been easier. Explain the viz helps users understand what’s important in a new dashboard.
- Ask Data: Developers can now use Ask Data directly from their desktop, which is a huge win. Developers don’t spend a lot time on the server/online, but by integrating Ask Data into the desktop, you’ll find a lot more data people using and benefiting from Ask Data’s capabilities, which is usually underused.
3. Building a data culture is key to success
Data, tools, and technology are nothing without people. People make the culture, and culture makes all the difference. A modern culture of data is one of experimentation and innovation, where people can accelerate business outcomes with rapid insights.
Focus on developing the mindsets and skills that drive action. Empower your people to uncover insights, ask questions, think critically, and challenge conclusions with data and experimentation. Here are some tips for building a data culture:
- Talk data constantly. Become comfortable with data through integrating data into daily work and conversations. Use data as the starting point for meeting agendas.
- Meet them where they are. Talk at the audience’s level of understanding when it comes to data. Don’t talk logistic regressions with folks who are not familiar with it. That will only further alienate and damage data culture. If your team doesn’t have a baseline of data literacy, make that a priority before focusing on data governance.
- Share openly. Share data where everyone can see it and have a baseline conversation about it.
- Let end users extract data. Allow analytics units to access and extract the data that they themselves have cleaned and QA’d to make incredible visualizations. It’s a win-win for breaking down data silos and ensuring better data accuracy within an organization.
- Inspire others with a vision. Paint a picture of what your organization can and should do with data. If you can't show the ways that you could solve problems or achieve success with data, others won’t invest the time/energy/resources into building a data culture.
We’re excited to see how the data community takes these enhancements, lessons, and opportunities to innovate even further in 2022. By using data viz ethically and responsibly, there’s no limit to the solutions we can create and good we can do in the world.
