Career Growth with Tableau

Dana launched into her presentation with a catchy subtitle: "From cubicle monkey to analytics guru in three steps (made easy with Tableau)." She presented the challenges and rewards of growing a Tableau user base across Wells Fargo, growing her professional network in the process. Her successful techniques for hosting group problem-solving and brainstorming have given her the confidence and skills to take Tableau dashboards directly to the CEO. In this respect, I'm reminded of the success story Dr. Jon Nakamoto presented at last year's customer conference, and this represents a trend of discarding stale Powerpoint slides in favor of live, rich analytics with Tableau. Of course, Dav Lion presented a technique during Devs-On-Stage for embedding Tableau within Powerpoint when slides are inescapable.

Dana's success story begins with last year's customer conference, where she met another Tableau user from Wells Fargo. Their early successes within their respective teams led to requests for demos by managers from multiple business lines, and they built enough interest to launch a user group. Dana has been invited to attend executive business strategy meetings, and Tableau has helped grow her network into Wachovia as the two companies merged.

How to use Tableau to get your analysis out of PowerPoint and into action

Turning to the topic of group problem solving, Dana described the key points for successful brainstorming sessions using Tableau.

• Sessions should involve 2-5 people
• Expect to need an overhead projector
• Make popcorn! (or other techniques to make folks comfortable)
• QA your data: ensure that it's clean, and do some preliminary analysis. Dana recommends preparing around five worksheets in advance. This provides a launching point for discussion, and gives the meeting leader some familiarity with the data.
• Schedule as 50-60 minute meeting
• Create an agenda and stick to it

Identifying outliers is a key goal of Dana's sessions, and she recommends a few techniques for glancing at the data from a different perspective:

• Table-calcs, such as % change, help expose trends that might otherwise be hidden.
• A "Q-tip" diagram is a compact view similar to a sparklines, but uses line width and color encodings to draw attention to outliers across a dimension such as time.
• An x-y scatterplot with size & color encoding for key dimensions are good at clustering related data points, leaving outliers as isolated marks.