Watson Does Mail and Analytics to Complement Inventing Recipes

November 19, 2014

IBM is beating the drum for Watson. “IBM Brings Watson Tinged Analytics to New Mail and Social Platform” reports about “an enterprise social collaboration platform with built in analytics.”

When I read the article, I thought of Semandex. My recollection is that this New Jersey-based company has a similar system. Perhaps the IBM collaboration function will be different from what Semandex offers.

My reaction to the flow of Watson “news” is that IBM is going to have to shift into high gear in order to generate $1 billion in revenue from scripts and open source software. With the $10 billion target looming 60 months out, I would suggest that IBM needs to make big sales to high profile clients quickly and in a serial fashion.

Right now Watson is enriching public relations and marketing types. IBM needs big, high margin sales. We have identified 36 companies providing more advanced functions than Watson. Time may be running out, particularly if an IBM competitor snaps up two or three of the outfits on our watch list.

Stephen E Arnold, November 19, 2014

RapidMiner Cloud Includes Wisdom of the Crowds Operator Recommendations

November 13, 2014

The article on Inside BigData titled RapidMiner Moves Predictive Analytics, Data Mining and Machine Learning into the Cloud promotes RapidMiner Cloud, the recently announced tool for business analysts. The technology allows for users to leverage over 300 cloud platforms such as Amazon, Twitter and Dropbox at an affordable price ($39/month.) The article quotes RapidMiner CEO Ingo Mierswa, who emphasized the “single click” necessary for users to gain important predictive analytics. The article says,

“RapidMiner understands the unique needs of today’s mobile workforce. RapidMiner Cloud includes connectors to cloud-based data sources that can be used on-premises and in the cloud with seamless transitioning between the two. This allows users to literally process Big Data at anytime and in any place, either working in the cloud or picking up where they left off when back in the office. This feature is especially important for mobile staff and consultants in the field.”

RapidMiner Cloud also contains the recently launched Wisdom of the Crowds Operator Recommendations, which culls insights into the analytics process from the millions of models created by members of the RapidMiner community. The article also suggests that RapidMiner is uniquely capable of integration with open-source solutions, rather than competing, the platform is more invested in source-code availability.

Chelsea Kerwin, November 13, 2014

Sponsored by ArnoldIT.com, developer of Augmentext

Enterprise Data Discovery for Self-Service Search

November 12, 2014

The article titled The Five Rules for Data Discovery on Computerworld discusses Enterprise Data Discovery. In the pursuit of fast-paced, accurate data analytics, Enterprise Data Discovery is touted in this article as a ramped up tool for accessing relevant information quickly. The first capability is “governed self-service discovery” which enables users to reformulate their data search on their own. This also allows for the blending of data types including social media and unstructured data. The article also emphasizes the importance of having a dialogue with the data,

“You also discovered that the spike in sales occurred in the middle of the media campaign and during the time of the spike, there was a major sporting event. This new clue prompts a new question – what could a sporting event have to do with the spike? Again, the data reveals its value by providing a new answer – one of the advertisements from the campaign got additional play at the event. Now, you have something solid to work on.”

According to the article, Enterprise Data Discovery offers a view of the road less travelled, enabling users to approach their discovery with new questions. Of course, the question that arises while reading this article is, who has time for this? The emphasis on self-service is interesting, but it also suggests that users will be spending a good chunk of time manipulating the data on their own.

Chelsea Kerwin, November 12, 2014

Sponsored by ArnoldIT.com, developer of Augmentext

eDigital Research and Lexalytics Team Up on Real Time Text Analytics

November 11, 2014

Through the News section of their website, eDigitalResearch announces a new partnership in, “eDigitalResearch Partner with Lexalytics on Real-Time Text Analytics Solution.” The two companies are integrating Lexalytics’ Salience analysis engine into eDigital’s HUB analysis and reporting interface. The write-up tells us:

“By utilising and integrating Lexalytics Salience text analysis engine into eDigitalResearch’s own HUB system, the partnership will provide clients with a real-time, secure solution for understanding what customers are saying across the globe. Able to analyse comments from survey responses to social media – in fact any form of free text – eDigitalResearch’s HUB Text Analytics will provide the power and platform to really delve deep into customer comments, monitor what is being said and alert brands and businesses of any emerging trends to help stay ahead of the competition.”

Based in Hampshire, U.K., eDigitalResearch likes to work closely with their clients to produce the best solution for each. The company began in 1999 with the launch of the eMysteryShopper, a novel concept at the time. As of this writing, eDigitalResearch is looking to hire a developer and senior developer (in case anyone here is interested.)

Founded in 2003, Lexalytics is proud to have brought the first sentiment analysis engine to market. Designed to integrate with third-party applications, their text analysis software is chugging along in the background at many data-related companies. Lexalytics is headquartered in Amherst, Massachusetts.

Cynthia Murrell, November 11, 2014

Sponsored by ArnoldIT.com, developer of Augmentext

Unicom Global Acquires Cognos

November 10, 2014

Catching up on old news: Just wanted to document that Cognos is no longer part of IBM. “Unicom Global Acquires Cognos Finance Business Analytics Software from IBM Corp.” The write up does not make clear exactly what Unicom acquired nor how much money IBM received in the deal. Presumably Unicom paid cash. Presumably IBM did not pay Unicom to take the Cognos bundle off its hands. For IBM analytics’ fans, you don’t need to worry. IBM owns SPSS and the dozens of systems and methods developed at its various research labs. Anyone remember Web Fountain?

Stephen E Arnold, November 10, 2014

Attensity: Downplaying Automated Collection and Analysis

November 7, 2014

I read “Do What I Mean, Not What I Say: The Text Analytics Paradox.” The write up made a comment which I found interesting; to wit:

Now, before you start worrying about robots replacing humans (relax—that’s at least a couple of years away), understand this: context and disambiguation within these billions of daily social posts, tweets, comments, and online surveys is they key to viable, business-relevant data. The way human use language is replete with nuance, idiomatic expressions, slang, typos, and of course, context. This underscores the magnitude of surfacing actionable intelligence in data for any industry.

Based on information my research team has collected, the notion of threat detection via automated collection and analysis of Internet-accessible information is quite advanced. In fact, some of the technology has been undergoing continuous refinement since the late 1990s. Rutgers University has been one of the academic outfits in the forefront of this approach to the paradox puzzling Attensity.

The more recent entrants in this important branch (perhaps an new redwood in the search forest) of information access are keeping a low profile. There is a promising venture funded company in Baltimore as well as a China-based firm operating from offices in Hong Kong. Neither of these companies has captured the imagination of traditional content processing vendors for three reasons:

First, the approach is not from traditional information retrieval methodologies.

Second, the companies generate most of their revenue from organizations demanding “quiet service.” (This means that when there is no marketing hoo hah, the most interesting companies are simply not visible to the casual, MBA inspired analyst.

Third, the outputs are of stunning utility. Information about quite particular subjects are presented without recourse to traditional human intermediated fiddling.

I want to float an idea: The next generation firms delivering state of the art solutions and have yet to hit the wall that requires the type of marketing that now characterizes some content processing efforts.

I am trying to figure out how to present these important but little known players. I will write about one in my next Info Today article. The challenge is that there are two dozen firms pushing “search” in a new and productive direction.

Stephen E Arnold, November 7, 2014

Textio is a Promising Text Analysis Startup

November 6, 2014

Here’s an interesting development from the world of text-processing technology. GeekWire reports, “Microsoft and Amazon Vets Form Textio, a New Startup Looking to Discover Patterns in Documents.” The new company expects to release its first product next spring. Writer John Cook tells us:

“Kieran Snyder, a linguistics expert who previously worked at Amazon and Microsoft’s Bing unit, and Jensen Harris, who spent 16 years at Microsoft, including stints running the user experience team for Windows 8, have a formed a new data visualization startup by the name of Textio.

“The Seattle company’s tagline: ‘Turn business text into insights.’ The emergence of the startup was first reported by Re/code, which noted that the Textio tool could be used by companies to scour job descriptions, performance reviews and other corporate HR documents to uncover unintended discrimination. In fact, Textio was formed after Snyder conducted research on gender bias in performance reviews in the tech industry.”

That is an interesting origin, especially amid the discussions about gender that currently suffuse the tech community. Textio sees much room for improvement in text analytics, and hopes to help clients reach insights beyond those competing platforms can divine. CEO Snyder’s doctorate and experience in linguistics and cognitive science should give the young company an edge in the competitive field.

Cynthia Murrell, November 06, 2014

Sponsored by ArnoldIT.com, developer of Augmentext

Predictive APIs: Will Search Vendors Play in This Sandbox?

October 31, 2014

I received a notice about new conference called “The First International Conference on Predictive APIs and Apps.” According to the write up I saw:

Several companies who are building predictive APIs and tools to make predictive app development easier will be at PAPIs (BigML, Datagami, Dataiku, Indico, Intuitics, GraphLab, Openscoring, PredictionIO, RapidMiner, Yhat). We’re expecting to see both actual and potential users who will share and learn how to use these products. Newcomers will learn and get inspiration from the keynotes, showcases and practical “predictive for all” user stories. Experts will also be interested in the sessions on technical challenges and in the panel discussion on the future of predictive APIs.

A number of search and content processing vendors suggest they deliver advanced analytics. Text analytics vendors are either feeding data into predictive engines or delivering outputs that are predictive.

Are predictive analytics one of the next big things? If so, traditional information retrieval and content processing companies are likely to be attending this conference on November 17 and 18, 2014.

At this time, IBM and Microsoft are on the program.

IBM will be addressing “intelligent APIs.” In the abstract for his talk, I did not see a reference to Watson. Microsoft’s talk abstract is not on the program page as of October 30, 2014.

Worth attending if you in the Barcelona area.

Stephen E Arnold, October 31, 2014

FirstRain Escapes “Death Spiral” Through Work of Penny Herscher

October 31, 2014

The article on Fortune titled The Company Was In a Death Spiral. She Brought It Back From the Brink lauds the work of Penny Herscher  at data analytics firm FirstRain. Herscher took over the company in 2004 after successful work at Cadence Design Systems, Simplex and Texas Instruments. FirstRain was a bankrupt company with a great prototype but no product. Herscher embraced the challenges posed by FirstRain and began her overhaul with a move from New York to California. The article goes on,

“She raised $20 million from new investors and hired a trusted team, including chief operating officer Y.Y. Lee, a mathematician and software engineer… Today, more than 50% of FirstRain’s senior leadership is women. The fledgling company had barely started developing a product when storms began brewing on the horizon. It was 2008. The global economy was beginning to collapse. “The wheels came off the bus,” Herscher says with lament. To survive, the company had to completely change course again…It pulled through.”

But only after major lay-offs and changes in the structure. Today FirstRain customers include IBM and Cisco, and it is only continuing to grow, with new offices in San Mateo. Herscher’s story of success is one of commitment and creative problem-solving.

Chelsea Kerwin, October 31, 2014

Sponsored by ArnoldIT.com, developer of Augmentext

Visualizing Data: Often Baffling

October 29, 2014

I read “72 Hours of #Gamergate.” I don’t follow the high buck world of video games. The write up contains oodles of data. Some of the information is in the form of bar charts. Other information is presented in words, spreadsheets, and graphics. I am okay with the bar charts. These have labels and numbers on the x and y axes. The visualization show below baffles me:

image

The image adds graphic impact. I have been in briefings in which senior executives and military brass have presented similar visualizations.

I suppose clarity is less important than sizzle. Analytics vendors, are you listening? I think not if this graphic is any indication of the way data are presented.

Stephen E Arnold, October 29, 2014

Next Page »