Watson Analytics Speed of Visualization
Dean Compher
21 December 2015
Quickly visualizing data you have in a file or a data warehouse can be very useful. If you are not familiar with your data, visualizing it can be a very time consuming task, but Watson Analytics can help you see if there is anything interesting in a file or database very quickly. Because it actually scans through your data and presents possible charts and graphs that answers questions you have not even asked yet, it can help you understand new data with little effort. You can also ask your own questions in a text format and have it draw graphs to provide answers, plus you can pick important fields in your data like purchase_total and have Watson Analytics show the fields that predict that data. You can even get started for free. This “speed of visualization” is very important when exploring new data.
In my previous Watson Analytics article, I discussed some of the basics and how to get started. In this one I’ll talk about a number of additional features, some of which are new since I wrote my last article. Some of these features are in the free version and some are not, so I’ll be sure to note this as I talk about the various features. This page shows the features in each edition and the price of each as well. Click the details link of each edition to see which features it includes. Some of the topics I discuss in this article are:
IMPORTANT NOTE!!! You must be signed into your Watson Analytics account to use the links below that point to instructions for particular features. You can see them in the free edition. If you are not logged in the first time you click one of these links you will be prompted for your user id and password. When you enter it you will land on the first page of the documentation. Click the link again in the browser session where you logged in and you should be taken directly into the instructions.
In the free edition you can upload your own delimited or Excel files and this has been there since Watson Analytics became available. Since my previous article many new sources of data have been added, like most major databases including DB2, Oracle and SQL Server. You just connect Watson Analytics to the source that interests you and create the data set without uploading files yourself. Here is the list of available sources as of the writing of this document:
More sources are sure to be added as time goes on. Please note that some of these are not traditional databases and include various Hadoop SQL sources like Hive and Impala. These sources can be in the cloud or on your own premises as long as your network will allow outside connections. You can even use existing Cognos reports a source of data for your Watson Analytics data set.
To connect to these various data sources, you must first create the connection to the data source and then you create the data set using the connection you created. Please see the instructions for creating data connections and then adding the data sets. Please note that when adding relational data sources you can filter rows and shape columns using SQL or just bring in whole tables.
As I noted earlier you can create Watson Analytics data sets by uploading data from your workstation. You can now connect Watson Analytics to cloud storage facilities like Box, Dropbox and others and point it at files in those services directly. This is available in the free and paid editions. These are just additional choices you have when you use the add data function as described in my previous Watson Analytics article.
To add data, click the add button (a plus sign in a circle that is in about the middle of your “Welcome” window) and then click the upload button in the pop-up presented:
If you want to try out Watson Analytics, but don’t want to bother with finding and uploading your own data, you can use sample data provided. All editions including the free one have the sample data ready to import into your own data set and explore. To use the sample data, just click the add button as above, except instead of clicking the Upload data button, choose the “Sample data” button next to it. You will be presented with several samples including a Campaign Effectiveness choice and several others.
You can share your visualizations (charts, predictions, etc.) you create in Watson Analytics from the free edition and you can share links and workspaces with others on your professional edition account. Before I jump into sharing visualizations it is important to talk about the Assemble feature first. It is one of the main functions you can do in Watson Analytics along with Explore and Predict. As you “explore” and “predict” your data you can “pin” various charts, graphs and predictions that you create when using the Explore and Predict functions. At some point you can use those pinned objects and put them together into a report or dashboard using the Assemble function. I introduced this topic here because you may want to share multiple objects at once and the Assemble function is one way of doing that.
I’ll start by taking about sharing abilities you have in the free edition. All of these ways involve putting the charts, graphs, assembled items, and other visualizations into a file and sending them to someone else. Therefore, the person receiving the files doesn’t even need Watson Analytics to view them. The ways that you can share are through e-mail, social media (as a Tweet or Post) or by just downloading the file or files. All of these ways of sharing allow you get the images in PNG image files, but with the email and download sharing methods you can also create them in a presentation or PDF. Since you would frequently use the visualizations in a presentation having them popped right into a presentation is quite handy. You can read about how this is done in the Share Visualizations page.
The next way of sharing a visualization is to share a link to it. The he recipient must be on the same Watson Analytics account with you to be able to use the link. With the link they can see the visualization you created. You just create the link within the chart or graph you are working on and then paste it into an e-mail or other application and send it to your colleague. You can read about how to share a link on the Sharing visualizations through a link page.
With the professional edition you can have two types of workspaces where you keep files and create charts, graphs, and other objects. These are the Personal Workspace and the Shared Workspace. As the names suggest only you can see the objects in your personal workspace and others can see objects in the shared workspace. When you put something in the shared workspace you can configure exactly who on the account can see it. Read more about this on the Share and move tiles page. The newest feature in sharing is the ability to add comments, have conversations about the data and poll other users about what they think of the various questions being asked of Watson Analytics. You can then search and filter these comments. This feature is explained on the Conversations page.
One of the main activities you can do with your data along with Explore, Predict and Assemble features is the Refine function. This function allows you to get your data into a format that is more conducive to your purposes. Once you make changes to the data set and save it, a new data set will be created with a new name of your choosing. The old data set will still be there in the original format, so don’t worry too much about making mistakes. All of these refinement features are in the free version.
One of the more simple and useful items renaming a column. Many times the data set you got will have names that are not very descriptive. To rename a column just bring up the data set in the refine function, right click on the column you want to rename it and select the “Change name” option. While you are in this dialog box you can also view the data type. To move a column just drag it to the new location.
You can add a filter on one or more columns so that you can eliminate rows that are not helpful to all users. So with this you can create different sets of data for different user classes, but only have to upload one set of data. You can also create new columns that are based on calculations of other columns, such as SUM and MEDIAN. You can also shape your data by adding hierarchies and groups of data.
***
Since this tool is so easy to use I suggest that you find a file of information that interests you and get started right away. It does not matter whether you are using business or personal data. You will not regret it. I am really interested in hearing the fascinating things that you find. Please post them to my Facebook Page or db2Dean and Friends Community.