2. 2
IBM Watson - Pros & Cons
Pros & Cons
Ease of use, sharing and collaboration
Descriptive, Diagnostic, Predictive and
Prescriptive analysis
Pattern discovery technology
Remove Hypothesis based analysis
No need to write mathematical models
and algorithm
Models can be exported as R Code
Models can be scheduled at a specified
time day /week/month etc..
Analysis can be exported to word,
powerpoint and HTML
Add commentary to the analysis:
Explains the key insights, focuses the
user's attention on the crucial details,
and recommends additional graphs that
the user should see to better understand
related patterns and the overall context
Pros Cons
It takes a dataset (max 12 columns) –
May be trail version limitation
The number of rows 10 million (Whether
is in a single table or split across
multiple relational tables)
The number of columns 500
Does not understand relational
databases
First we have to load the data to Watson
We have separate tools to load the data
to Watson (For end user to load the data)
Does not do the analysis to the level that
BTB does
Limited functions and features
Can’t use user defined functions as there
are in R
3. 3
IBM Watson - Pros & Cons
Pros & Cons
Access, blend, transform data from
multiple data sources
Automated Self-learning Data Cleansing
Draw any graph you know you want to
see (hypothesis driven)
Guided Analysis shows related graphs
and explains hidden patterns
Automated Statistical Validation of each
graph
Truly Dynamic Dashboards detect and
explain the key changes today
Privacy By Default leveraging K-
anonymity
Audit and monitoring of data / analysis /
model / collaboration
Automated K-fold Model Validation and
model simplification
It also has point-and-click data join,
transformation, and cleaning features
Pros Cons
It takes a dataset (max 12 columns)
4. 4
Datawatch and their Monarch solution
Pros & Cons
Simple to use. Business users are able to prepare all types of data by working directly with it on
premise, this is self service data prep solution
Automatic data extraction. Drag and drop, no scripting skills required.
Web page data extraction. Capture just the data without other page “noise”
Instant understanding. Visually navigate and filter data and metadata of any size
Multiple source connectivity. Out-of-the-box connectivity tools link to all major relational databases,
Hadoop, NoSQL, Salesforce.com and more.
Prep visibility. Automatically capture every change made for transparency.
Process automation. Save prep steps for reuse and reapplication
Powerful algorithms. Easily join disparate data without being a data scientist
Report consolidation. Append tables with a single mouse click
Watson-ready. Prepped data exports directly into IBM Watson Analytics, ready for analysis
BI integration: Connectors included to a variety of BI platforms including Oracle BI, Qlik and Tableau
allowing data to be blended without worrying where the data resides.
On premise: No need to upload data to the cloud, DW Monarch allows data blending/shaping to stay on
premise and then when ready a one time upload to IBM Watson Analytics can be initiated.
Low cost: A single user or DW Monarch is $500 per annum.