I've always been disappointed in Tableau for not having a built-in calendar view. Fortunately it's remarkably easy to construct one yourself using Tableau's wide array of date functions. All you need to get started is a data source with at least one date field.
As it turns out, this knowledge came in handy for a personal project I've been working on. I recently moved into the city from the far-flung suburbs of Virginia, and have been noticing my wallet feels a little lighter than usual. Even taking into consideration the usual suspects (higher rents & grocery costs) it didn't seem like I should be seeing quite this much damage to my bottom line.
I decided what I really needed was two things: A quick visual to let me identify days of above-average spending, and a breakdown of category spending by year and quarter. Below is what I came up with using data from one of my credit cards (lowered by a random factor for privacy reasons). If you want to learn how to set up your very own calendar control, read on after the jump.
As a resident of the DC area, I often find myself seeking cooler climes in order to escape the oppressive summer heat. The problem I always run into is that, while there are thousands of websites that'll tell you what the weather is today, very few give you a picture of a city's overall climate. When they do, they tend to present this information in such a way that makes comparisons across cities very difficult (weather.com, I'm looking at you). What if there were a way to summarize the climate for every major US city within a single view?
I took a stab at the question using free data from NOAA. Here's what I came up with. Click on a city to see the average temperature for each month.
Topics Covered: Creating a flowchart, Custom shapes, Formatting numbers within string labels
A colleague recently challenged me to create an interactive, automatically-updated version of his user acquisition flowchart in Tableau. The chart in question was a simple diagram of a website's user registration funnel, and he had been updating it manually for weeks at a considerable cost to his sanity. The product team laid out the following requirements:
The funnel needed to represent four basic steps. Without going into detail, we'll call them "landing", "sign-in page 1", "sign-in page 2", and "complete."
Three absolute metrics needed to be displayed for each of these stages: Users, estimated user value in dollars, and visits. Conversion rates also needed to be calculated for each (so we could say 25% of users went from "landing" to "sign-in page 1", 10% from "sign-in page 1" to "sign-in page 2", etc).
The graphic should allow for at-a-glance diagnosis of under-performing pages.
All metrics, both absolutes and conversion rates, should be able to be trended over time.
While I can't share the actual dashboard I designed for this project, below is one I put together from scratch based on a traditional eCommerce checkout funnel. Try out the filters and click on the icons to see trended views of each metric. More details on construction behind the jump.
