Editor's Note: IBM Watson Analytics is no longer available for purchase, as of July 31, 2018.
IBM Watson Analytics (which begins at $30 per user per year for the Plus edition) has gained wide recognition in the self-service business intelligence (BI) space thanks in part to an ever-present marketing campaign.
All of the buzz is deserved because this application is a winner with its technical capabilities.
Its highly advanced analytics engine works with a stellar natural language querying platform to make BI as easy to learn as it is powerful.
When you start using IBM Watson Analytics, it's clear that the machine learning (ML) and artificial intelligence (AI) behind it is an impressive accomplishment.
The assistant serves the roles of both servant and guide for users with varying levels of data science and data visualization expertise.
IBM Watson Analytics gets our Editors' Choice designation along with Microsoft Power BI and Tableau Desktop.
The guidance prompts in IBM Watson Analytics are very helpful for users who are new to analytics and to seasoned business analysts who would rather focus on getting rapid feeds of insights rather than on how the data is cooked.
Yet the tool enables those with deep data science skills to skip past the prompts and move directly to deep dives and decision trees.
It's a fine line to serve users on the entire spectrum of skills well, but IBM Watson Analytics pulls it off better than expected.
There are 32 connectors to ease use of data from those sources.
A sample listing of business connectors includes spreadsheets (CSV, XLS, TXT), Eventbrite, Hubspot, OneDrive, Paypal, SugarCRM, SurveyMonkey, and Twitter.
It's interesting that IBM Watson Analytics lists a Twitter connector for use, but users must access IBM Watson Social Media for the rest of the social tools.
This is where I found numerous connections to all social media (including Facebook, of course) as well as to blogs, comments, forums, and videos throughout the social media realm.
IBM Watson Analytics also lets you directly query a variety of databases including Cloudera Impala, Microsoft Azure ($14,300.00 at Microsoft Azure) , MySQL, Oracle, PostgreSQL, PostgreSQL on Compose, Structured Query Language (SQL) Server, Sybase, Sybase IQ, and Teradata.
Getting Started
While this is a full self-service BI tool, IT can lend a hand in loading company data to simplify things further for business users and to also ensure compliance and security rules are set as IT would have them.
However, users can easily load data themselves providing they have the credentials to access it.
Users are then prompted on setting rules.
Users can also shape and cleanse the data prior to uploading it by clicking the "Shape Before" button rather than the "Upload Now" button, which are both choices after clicking "+ New data" (the Add Data button) on the IBM Watson Analytics homepage.
Users can depersonalize customer names and other personally identifiable information (PII) to comply with privacy rules in the "Shape Before" function prior to upload.
I used the "Shape Before" button to move through the data and delete empty rows, but you can also clean up addresses and dates and other information to achieve conformity which aids in the analysis.
In other words, "Shape Before" lets user easily clean and prep the data with just a few clicks prior to running an analysis on it.
The system prompts the user with suggested cleaning and prep tasks that particular data set appears to the system to need.
The simplicity in this makes it easily undervalued when you consider that data prep typically takes more time than any other step in the analytics processes—yet here it is a matter of prompts and clicks and you're done.
After all data sets are added by the user or IT department, the user interface (UI) shows the data sets in the user's choice of icons or a table.
Either way, the data sets shown there can come from user uploads, data accessed through one or more IBM Watson Analytics' connectors, data sets that have been combined by the user within the system, or data sets from IBM Watson Social Media (a social media analytics tool accessible from within the IBM Watson Analytics tool).
Note that IBM Watson Analytics does not do streaming analytics; such is common with the Internet of Things (IoT) data and other sources and systems where information is streamed and analysis must be instant.
However, in IBM Watson Analytics, data can be refreshed frequently, as often as every 5 seconds or so, for a near-real-time read.
That will suffice for many use cases and certainly for the use cases IBM is aiming for: visualization, patterning, and social media point-in-time analysis.
It cannot be used for any use case that requires actual real-time, streaming analysis.
That may be a major shortcoming for some users.
In short, I found the data setup to be simple and straightforward and well-suited to the typical business analyst's skill level as well as within the reach of most other business users.
The system is highly intuitive, but it can be intimidating to beginners nonetheless.
Regardless of skill level, I highly recommend viewing the tutorials first.
Although most users will be able to poke and feel their way through the tool just fine, the tutorial will get you up and running faster, particularly if you haven't used a self-service BI tool before.
If you are experienced in data wrangling and analytics, then it's still a great time-saver to view the tutorials initially.
The Discovery Process
The user can move to the discovery phase (that is, running the analytics) in one of two ways: either by clicking on a specific data set or by typing in a question in the question bar.
Clicking a specific data set provides prompts called "Starting Points," which deliver insights IBM Watson Analytics has calculated to likely be of the most interest to you.
Yes, that means instant insights delivered via commonly used and prebuilt algorithms.
If you type a question into the question bar instead, then the system looks for that information in all of the data sets you have added/loaded rather than in one selected data set.
You can also click the Help tool to learn how to pose a question to the system.
I found that querying this tool works very well with natural language, meaning I could type in a question as I would ask a colleague and then get insights and suggestions that fit.
I could also type in a series of key words and get back the same thing.
For example, I typed "What types of products are more popular right around the Black Friday period?" and then typed as a separate question "products, November." In answer to both, the system pulled up several suggested questions and data sets in order of relevancy.
I could then retrieve insights by clicking the suggestion that best fit.
I could also explore each suggested starting point at my leisure wherein I would find visualized insights and a scrollable infographic for rapid consumption.
You can even change the visualization form with a couple of clicks.
It's here in the visualization drill-down that I stumbled a bit and why I recommend watching the tutorials first.
I thought I had to go back and ask another question, to repeat the entire process, just to add another factor (input) to my question.
It was only after a company rep walked me through the demo that I learned I could do that via the "Color Value" function—which is found in a bar at the bottom of the visualization.
Big "duh moment" there, but it reminded me that even though IBM Watson Analytics is highly intuitive, it's a lot like getting a new mobile phone and awkwardly figuring out the commands only to wonder later why you found any of it difficult.
Yeah, this was exactly like that.
Intuitive and easy but only after I did some head scratching and mumbled aloud, "Wonder where they put the tool to do XYZ?" Still, compared to most of its competitors, IBM Watson Analytics is a breeze to use.
For more complex querying, type questions that begin with, "What drives X" or "What predicts X," which will take you to much deeper dives and intricate decision trees.
I found this to be highly effective in discovering insights that might otherwise have taken me much longer to even figure out what to ask of the data.
The Display function—the third step in this system's process—lets you choose visualizations, save them, and/or publish to a dashboard or an infographic that you build in a couple of clicks or drag-and-drops.
It is designed to let you easily share the insights you've found with others so you can collaborate on actions to take next, or on expansions or refinements to be made in the analysis.
Pricing and Versions
There is both a free trial and a free edition with limited features available.
Those that are testing the system are automatically steered to the trial edition, which offers 1 megabyte (MB) of storage.
Users can decide to purchase or move to the more limited free edition at the end of the trial period.
The Plus edition, the one I tested, is $30 per month per user and is essentially the free trial edition with 2 gigabytes (GB) of storage, along with the added bonus of being able to load larger data sets.
The Professional edition is designed for enterprise use and can accommodate multiple users, more complex collaborations, and more data connectors than the other editions.
The cost for the Professional edition is $80 per user per month or $960 per user per year with 100GB of storage.
IBM Watson Analytics provides insights that can help businesses in various industries from retail to health care.
There's no doubt that detecting hidden trends and correlations in structured and unstructured data can be valuable.
Companies can analyze the circumstances driving business events and use the data to inform future decisions.
Overall, IBM Watson Analytics is beautifully designed to deliver a highly intuitive UI and a simplified user experience (UX).
The prompts are indeed intelligent and guide even beginners in a smart direction, which will go far in companies aiming for companywide data democratization, which is almost everyone in the organization using data analytics in their work.
It's also very smooth at moving users to deeper insights even if their data science skills are somewhat limited.
Yet, highly experienced data scientists can skip the prompts and go straight to complex querying and decision trees and avoid most of the frustration associated with "user-friendly" software.
That is not to say that IBM Watson Analytics is perfect or suited for all use cases.
It's inability to handle streaming data and analytics on the fly will be a major drawback to some (but a non-issue to many others).
If near-real-time is sufficient for your use cases and your employees have a mix of skill levels, then you'll find this tool is more than up to the task.
Pros
Accessible user interface.
Smart guidance features.
Impressively fast analytics.
Robust natural language querying.
View More
The Bottom Line
IBM Watson Analytics is an exceptional business intelligence (BI) app that offers a strong analytics engine along with an excellent natural language querying tool.
This is one of the best BI platforms you'll find and easily takes our Editors' Choice honor.
Editor's Note: IBM Watson Analytics is no longer available for purchase, as of July 31, 2018.
IBM Watson Analytics (which begins at $30 per user per year for the Plus edition) has gained wide recognition in the self-service business intelligence (BI) space thanks in part to an ever-present marketing campaign.
All of the buzz is deserved because this application is a winner with its technical capabilities.
Its highly advanced analytics engine works with a stellar natural language querying platform to make BI as easy to learn as it is powerful.
When you start using IBM Watson Analytics, it's clear that the machine learning (ML) and artificial intelligence (AI) behind it is an impressive accomplishment.
The assistant serves the roles of both servant and guide for users with varying levels of data science and data visualization expertise.
IBM Watson Analytics gets our Editors' Choice designation along with Microsoft Power BI and Tableau Desktop.
The guidance prompts in IBM Watson Analytics are very helpful for users who are new to analytics and to seasoned business analysts who would rather focus on getting rapid feeds of insights rather than on how the data is cooked.
Yet the tool enables those with deep data science skills to skip past the prompts and move directly to deep dives and decision trees.
It's a fine line to serve users on the entire spectrum of skills well, but IBM Watson Analytics pulls it off better than expected.
There are 32 connectors to ease use of data from those sources.
A sample listing of business connectors includes spreadsheets (CSV, XLS, TXT), Eventbrite, Hubspot, OneDrive, Paypal, SugarCRM, SurveyMonkey, and Twitter.
It's interesting that IBM Watson Analytics lists a Twitter connector for use, but users must access IBM Watson Social Media for the rest of the social tools.
This is where I found numerous connections to all social media (including Facebook, of course) as well as to blogs, comments, forums, and videos throughout the social media realm.
IBM Watson Analytics also lets you directly query a variety of databases including Cloudera Impala, Microsoft Azure ($14,300.00 at Microsoft Azure) , MySQL, Oracle, PostgreSQL, PostgreSQL on Compose, Structured Query Language (SQL) Server, Sybase, Sybase IQ, and Teradata.
Getting Started
While this is a full self-service BI tool, IT can lend a hand in loading company data to simplify things further for business users and to also ensure compliance and security rules are set as IT would have them.
However, users can easily load data themselves providing they have the credentials to access it.
Users are then prompted on setting rules.
Users can also shape and cleanse the data prior to uploading it by clicking the "Shape Before" button rather than the "Upload Now" button, which are both choices after clicking "+ New data" (the Add Data button) on the IBM Watson Analytics homepage.
Users can depersonalize customer names and other personally identifiable information (PII) to comply with privacy rules in the "Shape Before" function prior to upload.
I used the "Shape Before" button to move through the data and delete empty rows, but you can also clean up addresses and dates and other information to achieve conformity which aids in the analysis.
In other words, "Shape Before" lets user easily clean and prep the data with just a few clicks prior to running an analysis on it.
The system prompts the user with suggested cleaning and prep tasks that particular data set appears to the system to need.
The simplicity in this makes it easily undervalued when you consider that data prep typically takes more time than any other step in the analytics processes—yet here it is a matter of prompts and clicks and you're done.
After all data sets are added by the user or IT department, the user interface (UI) shows the data sets in the user's choice of icons or a table.
Either way, the data sets shown there can come from user uploads, data accessed through one or more IBM Watson Analytics' connectors, data sets that have been combined by the user within the system, or data sets from IBM Watson Social Media (a social media analytics tool accessible from within the IBM Watson Analytics tool).
Note that IBM Watson Analytics does not do streaming analytics; such is common with the Internet of Things (IoT) data and other sources and systems where information is streamed and analysis must be instant.
However, in IBM Watson Analytics, data can be refreshed frequently, as often as every 5 seconds or so, for a near-real-time read.
That will suffice for many use cases and certainly for the use cases IBM is aiming for: visualization, patterning, and social media point-in-time analysis.
It cannot be used for any use case that requires actual real-time, streaming analysis.
That may be a major shortcoming for some users.
In short, I found the data setup to be simple and straightforward and well-suited to the typical business analyst's skill level as well as within the reach of most other business users.
The system is highly intuitive, but it can be intimidating to beginners nonetheless.
Regardless of skill level, I highly recommend viewing the tutorials first.
Although most users will be able to poke and feel their way through the tool just fine, the tutorial will get you up and running faster, particularly if you haven't used a self-service BI tool before.
If you are experienced in data wrangling and analytics, then it's still a great time-saver to view the tutorials initially.
The Discovery Process
The user can move to the discovery phase (that is, running the analytics) in one of two ways: either by clicking on a specific data set or by typing in a question in the question bar.
Clicking a specific data set provides prompts called "Starting Points," which deliver insights IBM Watson Analytics has calculated to likely be of the most interest to you.
Yes, that means instant insights delivered via commonly used and prebuilt algorithms.
If you type a question into the question bar instead, then the system looks for that information in all of the data sets you have added/loaded rather than in one selected data set.
You can also click the Help tool to learn how to pose a question to the system.
I found that querying this tool works very well with natural language, meaning I could type in a question as I would ask a colleague and then get insights and suggestions that fit.
I could also type in a series of key words and get back the same thing.
For example, I typed "What types of products are more popular right around the Black Friday period?" and then typed as a separate question "products, November." In answer to both, the system pulled up several suggested questions and data sets in order of relevancy.
I could then retrieve insights by clicking the suggestion that best fit.
I could also explore each suggested starting point at my leisure wherein I would find visualized insights and a scrollable infographic for rapid consumption.
You can even change the visualization form with a couple of clicks.
It's here in the visualization drill-down that I stumbled a bit and why I recommend watching the tutorials first.
I thought I had to go back and ask another question, to repeat the entire process, just to add another factor (input) to my question.
It was only after a company rep walked me through the demo that I learned I could do that via the "Color Value" function—which is found in a bar at the bottom of the visualization.
Big "duh moment" there, but it reminded me that even though IBM Watson Analytics is highly intuitive, it's a lot like getting a new mobile phone and awkwardly figuring out the commands only to wonder later why you found any of it difficult.
Yeah, this was exactly like that.
Intuitive and easy but only after I did some head scratching and mumbled aloud, "Wonder where they put the tool to do XYZ?" Still, compared to most of its competitors, IBM Watson Analytics is a breeze to use.
For more complex querying, type questions that begin with, "What drives X" or "What predicts X," which will take you to much deeper dives and intricate decision trees.
I found this to be highly effective in discovering insights that might otherwise have taken me much longer to even figure out what to ask of the data.
The Display function—the third step in this system's process—lets you choose visualizations, save them, and/or publish to a dashboard or an infographic that you build in a couple of clicks or drag-and-drops.
It is designed to let you easily share the insights you've found with others so you can collaborate on actions to take next, or on expansions or refinements to be made in the analysis.
Pricing and Versions
There is both a free trial and a free edition with limited features available.
Those that are testing the system are automatically steered to the trial edition, which offers 1 megabyte (MB) of storage.
Users can decide to purchase or move to the more limited free edition at the end of the trial period.
The Plus edition, the one I tested, is $30 per month per user and is essentially the free trial edition with 2 gigabytes (GB) of storage, along with the added bonus of being able to load larger data sets.
The Professional edition is designed for enterprise use and can accommodate multiple users, more complex collaborations, and more data connectors than the other editions.
The cost for the Professional edition is $80 per user per month or $960 per user per year with 100GB of storage.
IBM Watson Analytics provides insights that can help businesses in various industries from retail to health care.
There's no doubt that detecting hidden trends and correlations in structured and unstructured data can be valuable.
Companies can analyze the circumstances driving business events and use the data to inform future decisions.
Overall, IBM Watson Analytics is beautifully designed to deliver a highly intuitive UI and a simplified user experience (UX).
The prompts are indeed intelligent and guide even beginners in a smart direction, which will go far in companies aiming for companywide data democratization, which is almost everyone in the organization using data analytics in their work.
It's also very smooth at moving users to deeper insights even if their data science skills are somewhat limited.
Yet, highly experienced data scientists can skip the prompts and go straight to complex querying and decision trees and avoid most of the frustration associated with "user-friendly" software.
That is not to say that IBM Watson Analytics is perfect or suited for all use cases.
It's inability to handle streaming data and analytics on the fly will be a major drawback to some (but a non-issue to many others).
If near-real-time is sufficient for your use cases and your employees have a mix of skill levels, then you'll find this tool is more than up to the task.
Pros
Accessible user interface.
Smart guidance features.
Impressively fast analytics.
Robust natural language querying.
View More
The Bottom Line
IBM Watson Analytics is an exceptional business intelligence (BI) app that offers a strong analytics engine along with an excellent natural language querying tool.
This is one of the best BI platforms you'll find and easily takes our Editors' Choice honor.
