Not only is Tableau a great data analysis tool, it’s also really useful for data discovery. One of the helpful ways you can use Tableau is to uncover inconsistencies and holes in your data structure. I do this by finding out what data unexpectedly has null values.
I recently got an email from a former student explaining that they were trying to recreate something similar to the image below in Tableau.
The trouble she was running into was recreating the total header called “Planned Cost by Perf”. If you try to use a measure as a discrete header, you end up with the value summed for each partition. Here is an example using Superstore data:
When calculating growth rates from one date period to another, it’s important to compare apples to apples. For instance, when building a graph in Tableau to compare quarterly sunscreen sales in Seattle, I probably wouldn’t want to compare Q3 Sales (July - September) to Q2 Sales (April - June) because there will be more sales in Q3. The product has a cyclical sales cycle. Instead, I would rather compare Q3 Sales of this year to Q3 Sales of last year to more accurately understand growth rates.
Comparing progress against a goal in Tableau is a common use case and I wanted to share a few tips about how I like to do this.
Tableau is usually the last 20% of any data communication effort. What is the first 80%? This document is meant to provide an example of what the finished output of data preparation for Tableau should look like.
After scouring the internet, I couldn’t find a good reference for beginners on how to prepare your data for Tableau so I figured, it looks like I’m going to have to do this myself! This is meant to be an introductory guide only, if you have more detailed questions feel free to send me an email.
Imagine trying to drive a Maserati at 180 miles per hour in traffic while keeping your entire focus on the rear view mirror. Impossible, right? Most data-driven reports are like this; aesthetically pleasing and historically-focused.
They are expensive and beautiful for a little while but almost always result in accidents.
How does OneNumber focus you on the windshield instead of the rear-view mirror? We study those historical results to inform what actions can improve future behavior.
American Airlines has a KPI called “D0”. It means no flight can depart late, no matter what or there are severe consequences.
“Many things go into whether or not a flight arrives on time. American’s management argues that:
“What they can most control is whether the flight departs on time. And if it does, that’s going to be the single biggest driver of on time arrivals. So American Airlines management is singularly focused on what they refer to as “D0” — departing exactly the minute that a flight is scheduled to depart (the government considers a flight to be ‘on time’ when it arrives within 15 minutes of schedule).”
Why do some bartenders get bigger tips than others? This was one of the first problems OneNumber sought to solve.
We interviewed bartenders, bar owners, managers and patrons. We asked them “what are the 3-5 things a bartender does that earns them greater tips?” We got over forty different answers from tattoos to revealing clothing.
We ran the responses through our algorithms and settled on just four things that a bartender can do *right now* to improve their tips. What do you think they are?