TDWI | Training & Research | Business Intelligence, Analytics, Big Data, Data Warehousing

Datawatch Revamps Monarch for Self-Service Data Prep

Datawatch's Monarch 13.3 release adds new analytical functions and algorithms. It also bundles new automated data loading capabilities, along with an Angoss export function.

• By Steve Swoyer
• May 16, 2016

Last month, veteran data integration (DI) software vendor Datawatch Inc. announced a version 13.3 of its Monarch self-service data preparation (data prep) product.

The new Monarch 13.3 adds new analytical functions and algorithms, starting with support for fuzzy matching. (It uses fuzzy logic to identify similarities or inconsistencies in different data sources.)

Also new is a feature Datawatch dubs "Load Plan Visualization." It provides a visual representation of a data-load plan, including its constitutive operations. It tracks data lineage, too.

Monarch 13.3 gets a logic refresh thanks to new if-then-else logic; new logic designed to improve data cleansing; and new fiscal-period logic. Monarch 13.3 performs what Datawatch calls "smart append operations," which -- in some cases -- can automate the integration of similar data values from different sources. Finally, Monarch can now export data flows for products from Angoss Software Corp. Angoss markets customer intelligence (KnowledgeREADER), data mining (KnowledgeSEEKER), and predictive analytics (KnowledgeSTUDIO) solutions, among others.

Datawatch's pitch with Monarch is a little different from that of its competitors. Vendors such as Alteryx Inc., Paxata Inc., and Trifacta Inc. tend to emphasize a similar use case: that of the analyst, data scientist, data engineer, or data explorer who wants to use a data set for analysis.

Datawatch emphasizes what it calls "operational" data prep. The focus in this context isn't on preparing a data set to support data discovery or advanced analytical experimentation but to address operational use cases ill served by the traditional, IT-provisioned DI model. They're low-value, one-off, or the business needs them more quickly than IT can deliver them.

This is a departure from orthodoxy. The case for self-service data prep is relatively straightforward. Self-serving analysts need data. Data must be prepared (i.e., integrated). IT-provisioned DI cannot address the needs of the self-serving analyst, who often prefers raw, uncleansed data, warts and all. IT-provisioned DI, by contrast, gives priority to data consistency and cleanliness. Self-service data prep is usually performed to support one-off or highly specialized use cases. IT-provisioned DI emphasizes the development and instantiation of repeatable, manageable, auditable data flows.

Both the experimental/exploratory and the operational data prep use cases have something in common: from the perspective of the self-serving data discoverer or the frustrated business person, IT-provisioned DI is insufficiently responsive -- it measures time-to-delivery in weeks or months, not hours or days -- and, moreover, is too tightly controlled.

A self-service data prep tool puts some degree of power -- autonomy and agency -- in their hands. Let's face it, preparing EBCDIC data set extracts from the mainframe, or CSV files from the legacy network fileserver, probably isn't as sexy as custom-tailoring a data set for experimental analysis.

It has a more immediate salience, however.

"From our perspective, we see data prep spanning both [use cases] and we see [operational data prep] as a much larger opportunity," Datawatch chief marketing officer Dan Potter told Upside.com in an interview last month. "We see it as a much larger piece of functionality that's not just going to be a feature in a BI tool. It's going to be useful in its own right."

[Editor's note: Previous posting mistakenly reported the new version number was 13.1.]