KDnuggets Home » News » 2016 » Feb » Software » IBM Watson Analytics for social media analysis ( 16:n08 )

IBM Watson Analytics for social media analysis


We explore IBM Watson Analytics features and what it can do with your data set. The IBM Watson Anlaytics social media add on is available for preview until March 1st.

What would you do when you have a critical decision to make & your data scientist or the business analyst is not available? Enter IBM Watson Analytics. Just Ask Watson! watson_analytics_ibm_features
IBM Cognitive Computing initiative is changing the way businesses are analysing data. You no longer need an IT expert or the tech guy to do complex data analysis. Watson Analytics offers a full range of cloud based self-service analytics, including access to easy-to-use data refinement & data warehousing services that make it easier for business managers to cleanse, prepare, analyze, visualize & as a last step present it creatively!

If you haven’t signed up for Watson Analytics for Social media analysis, here is the freemium version.
Let us explore the five core components of Watson Analytics namely - Explore, Predict, Assemble, Social Media, Refine. The Social media add-on being the recent addition.

Getting Started with Watson Analytics The freemium version is currently restricted to a maximum of 12MB and can contain no more than 50 columns. A dataset of KDnuggets twitter activity during the timeperiod 15th Dec 2015 -Jan 10th 2016 was uploaded to explore some insights about our twitter page. Check the data score for the quality of the data (a readiness rating for Watson analysis).

1. Refine - Organize your data  
Review and adjust your uploaded data to Watson Analytics. With Refine, you can rename columns, choose relevant columns, organize the data using groupings & hierarchies, view data scores(quality of data for analysis) & visualize data values, filter data & add new calculation columns to the dataset.

2. Explore - Interactive Q & A with Watson!
Not sure where to start? Don't worry IBM Watson provides you a starting point by suggesting initial analysis with relevant questions which you can simply click to explore. Explore is a powerful feature where you can type in your questions and get answers as instant visualizations. ibm-watson-analytics-project Alternatively, a user can select variables from the NLP Q & A box to see relevant questions. Here, we tried to ask questions like how does engagement rate compare by url clicks & likes? How does detail expands influence retweets? How does user profile clicks compare by day name?

3. Predict - Key predictors?
Predict is the tool that Watson provides to predict one or more target variable based on the others. It shows you the predictive strength & correlation between variables in your data. predictors_ibm_watson_analytics_project This corresponds to classification or regression depending on whether the target variable is categorical or continuous. We checked for what are the predictors of retweets, detail expands, user profile clicks and engagement rate etc.

4. Social Media - Discover social insights !
Enter keywords that you wish to track from the social sites. These keywords/topics pull data from a collection of internet communities, blogs, forums and online comments. As an example, below you see two topics for “KDnuggets” and “data scientist”. ibm-watson-analytics-social-media-add-on
When you run the social media analysis, you can view visualizations for your project’s Topics, Themes, as well as analytics on Sentiment, Geographic, Demographics, Active Authors and Sources.
We found interesting insights about our audience in terms of geography- Our active audience on twitter is from US, Canada, Australia, Brazil, India & South Africa. Some other country audience responding to our tweets within the Dec-Jan time period are Saudi Arabia, Chile, Algeria, & Peru. We were able to identify the top authors who are talking about us and the number of mentions. Click on any Topic to see topic suggestions displayed as word clouds of other keywords that also occur with your topic terms. You can exclude or include terms to make your data analysis more precise & suitable to your business scenario.

5. Assemble - Impress your audience with a great presentation
Once your insights are all saved into a collection. You can present them in a creative way as a timeline video or as reports. Assemble provides an interface to create "authored workbooks".

Business users, non-technical managers, domain experts can use information for any number of purposes depending on the business scenario. IBM Watson Analytics opens endless possibilities by giving any user a power to access insights!

Tell us how IBM Watson Analytics changed the way you derived insights for your project in the comments.