So I had a lofty goal to do this every day, but building product got in the way so I took a long break. As a result I will call this one Viz of the Month. This time I picked the Pie Chart (includes donut 🍩) because it is controversial in the data world. Most experts will say to never use it, but yet I see it in the wild all the time. My guess is people like the way it looks. So it gets points for aesthetics and therefore gets used.
Generally speaking pies fall apart when you have more than one measure and one dimension in play. Also that dimension better not be dense with lots of data points. Lastly, and generally speaking, a basic bar chart will often represent the data and the variations in the data better.
Here is what you don’t want to do:
So some rules for when to use a pie?
Small number of data points.
One dimension and one measure.
Sum of the data should be meaningful.
When you want to show a total with the parts. (Donut use case)
I say use it if you want but follow the simple rules above and some basic sense of visualization style. When looking at the charts above you can see they are way too dense to be meaningful at a glance. If you want to know more about the dreaded pie chart and why take a look at these link, but form your own option and choose wisely.
For mobile, because of touch, we are able to make them a bit more useful by allowing users to touch and interact with the pie to easily see the details about each slice. This style pie is available in Day by Day and Synopsis but we have a rule of less than 10 slices for our default.
Tag Cloud (aka Word Cloud) is a fun way to represent text or dimensional data. The size (font) or color of the data helps to easily visualize data. In the above example Tables are selling a lot but with high Shipping Cost. While at the same time Copiers and Fax (yes very old sample data 🤪) have hight Sales but Shipping Costs are low.
This visualization does have some drawbacks even though it looks really cool. Some of the drawbacks are the following:
Lots of text is hard to see.
Small variations in data may not be easily detectible.
Large words take up a bunch of space.
Same data is often better represented in another visualization.
Even with these drawbacks this is often a good chart to include in a dashboard or analysis you plan share with consumer users. As part of an overall analysis tag clouds can work well and can help draw users attention to data.
When loading this visualization as is on mobile devices they are responsive and interactive.
Also in our app Oracle Synopsis we have added some great features that work best on a device, like mini maps and zooming. This allows for some really dense Tag Clouds!
Boxplot is considered an exploratory chart that lays out data into quartiles with minimum and maximum values shown at the ends of “whiskers”. The whiskers are an indication of the variability of the data. Longer the whiskers the more variable the data. In the chart above, and with all categories on the x-axis, there is a 5 number summary including, minimum, first quartile, median (second quartile), third quartile, and the maximum.
Like the bar chart this can be represented vertically or horizontally, this choice is mostly for readability and might vary based on the density of the categories or the amount of space available for the visualization.
From a layout perspective this chart is a lot like a bar with a numeric value on the Values (Y-Axis) and some dimensional data on the Category (X-Axis). The additional data, in the above case, is Weekday and this shows the distribution of Sales across days of the week by Product Category.
At a glance this chart type is great for understating the data variations and the median for a given category, but is more useful when each category or data point is examined. The tooltip is a key tool for finding these details and in the above example, additional related data is added to the tooltips.
In general this is a great chart for the right data with the correct exploratory use case in mind. If the goal is simple value comparisons on small amounts of data with limited variation than the complexities of this chart will just get in the way of the analysis. For more explanation and details on a boxplot chart you can read this article. If you want to learn more and play with Oracle Analytics check out this free course on Udemy – https://www.udemy.com/augmented-analytics/.
This is basically the king of charts because it covers so many data data visualization needs. I will cover the basics and include vertical and horizontal but let’s leave stacked out of this summary.
Bar is best when comparing various categories of data; one axis of the chart shows the categories being compared while the other represents the metric. One of the reasons it is so great is because it is really simple to understand relative differences between categories. In my sample order data you can easily tell that West is selling the most and overall Toys are moving more than any other category (by almost double). When you sort the bars these relationships really stand out!
This chart type also works great in a trellis and works well when you want to group multiple categories in one chart. In my example below I choose color to represent category and we can now see that the West is the region driving a large portion of these Toy sales. This can also be seen in the trellis and further more changes across years can be seen. In the 2 chart example above we could not have come to the same conclusion.
When on your mobile device, and in this case with Oracle Analytics – Day by Day, you will see bar charts come up fairly often. Users can change the chart type easily but when searching for data on your device we find that many times the results can we well represented in a bar. For dense bars we allow smooth zooming and scrolling along with mini-maps.
At some point down the road I will go over pie charts but often times you will find switching any pie chart to a bar chart makes for better representation of data visually.
Correlation Matrix is a great way to show simple relationships between various metrics. It is very easy to use because all you need is a set of 2 or more metrics to make it useful. With this sample data we can easily see that Shipping Cost has a fairly strong negative correlation on Quantity Ordered and Profit. Might not be a revelation but it’s sample data, so hopefully you get the point.
Works great with filters and in a trellis allowing you to see the same relationships across various categories. Great use case for marketers or when analyzing sales data in general.
Also works well on mobile devices. All Oracle Analytics visualizations are responsive and adaptive out of the box.
Recently there has been a bunch of chatter around asking data and I wanted to share an old but great feature in Oracle Analytics. Around 2015 we launched home page search allowing users to search for answers to their data questions.
Right on our home page a user can simple ‘ask’ their data by using a familiar search interface. They can do this both structured and unstructured (speaking or freeform typing) and we just take care of the rest.
Structured search using tokens to get the data driving the question at hand.
Unstructured searching with voice or freeform typing leveraging NLP to get the same results.
As you can see it is really easy to ask your data and get quick results. Stay tuned for how to set this up and optimize your search experience for users leveraging some of the platforms great features like Data Sets, Advanced Data Preparation, Data Flows and more.
You can fully uninstall Oracle DV for Desktop on Mac by downloading this script and following the instructions below.
Right-click on the script and select Save As to save the file on your Mac.
Open Terminal and change the working directory to where the script was downloaded:
$ cd /location/of/file
Make the script executable and the run it with sudo:
$ chmod +x ./uninstall-dvdesktop.sh
$ sudo ./uninstall-dvdesktop.sh
Finally, delete the uninstall script.
Again keep in mind that this is a full uninstall of Oracle Data Visualization for Desktop including all artifacts like data sets and projects. In case the download does not work here is the contents of the script.
# Uninstall Data Visualization for Desktop on Mac echo “Uninstalling Data Visualization for Desktop on Mac…”
sudo rm -rf “/Applications/Oracle Data Visualization for Desktop.app” sudo rm -rf “/Applications/DVDesktop.app” sudo rm -rf “/Applications/Oracle Data Visualization Desktop Configure Python.app”
# Uninstall All Data and Projects echo “Uninstalling All Data and Projects…”
In older versions of Data Visualization each project launch, data set edit, or really anything would launch a new browser tab. Well this ending up leading to some bad issues including making it hard to find the right page, multiple instances of the same thing, and worst of all, closing tabs before you hit save. 😥
We called this browser tab hell, and we, like our end users, hated it!
So to fix this we started and internal project named tab-less navigation where we introduced an internal navigation flow for all of our key objects including:
In this model users can easily navigate back to where they were using our back button.
Also when a user wants to open content in a new tab they have that option.
Lastly for users that love having tons of tabs open and enjoyed tab hell they can just set it as their default.
So with the newest release or OAC the problem with tab hell is no more! I will admit there may be a random page or two the opens in a new tab but that is mostly in the area of administration, and we will clean that up over time. We hope you like the change and enjoy analytics without tab hell.
Day by Day hit the App Stores for iOS and Android today and there are a ton of great new features here are just a few.
Multiple Bring Backs
Now you can set multiple bring backs on each card and have the card return to the top of your Smart Feed for multiple reasons. Select ‘Bring Back’ on a card and simply tap Location, Schedule or Contacted by (Android only) option.
Bring Back when Contacted By (Android Only)
On Android, Go to the Bring Back screen and select the ‘Contact’ option which will walk you through selecting a contact. When that contact calls or texts you the card you set it on will surface on the top of the Smart Feed and notify you.
The filter within feature allows you to constrain the data from the search results allowing users to adjust the result to meet their exact need. It allows for the removal of dimensions, dimensions members, and any measure (fact) from the result. All with a simple and consistent user interface.
Manage Bring Backs
The Smart Feed contains a list of analytics that should be meaningful to the user based on what’s going on in the business and their day. Over time users may want to see and manage why a card shows up on the feed.
Simply go to the settings page and tap the Bring Backs to edit any currently active bring backs.
Popular Search Leaderboard
Over time Day by Day will proactively deliver analytics to users Smart Feed. The leaderboard is the first ‘system generated’ item that we push to user’s Smart Feeds.
In the latest release of OAC the team added a bunch of new visualizations and one of them has quickly become my favorite. It’s the correlation matrix!
This is my favorite because of the simplicity of it. Turns out I love visualizations that need no explanation! You just add a bunch of numbers to the visualization and there you have it. You can easily see how metrics correlate and how changes in one impact another. Hopefully leading to better business insights.
In case you want to get more details on some of the newest OAC features you can see them all here.