Data Visualization: Unusual and Unnecessary Terminology
March 19, 2019
I read “5 Reasons Why Data Visualization Fails.” Most of the information in the write up applies to a great many visualizations. I have seen some pretty crazy graphs in my 50 year career. A few stand out. The Autonomy heat maps. Wild and crazy radar maps. Multi axis charts which are often incomprehensible.
The problem is that point and click options present data. The “analyst” often picks a graph that keeps a general, a partner in a venture firm, or a group of rubes entranced.
The article touches upon other issues ranging from a failure to think about the audience to presenting complex visualizations.
I do have one major objection to the article. From my point of view, the “phrase data overload” or “large flows of information” express the concept of having a great deal of information. The article uses the phrase “data puking.” The phrase is unnecessary and off putting to me.
Stephen E Arnold, March 19, 2019
A Justification of Making Things Up?
March 13, 2019
I read “Gut Feelings Often Trump Real Data in Driving Business Decisions, Says Forrester.” The write up is interesting for several reasons. First, Forrester, like other mid tier consulting firms, generates reports about companies with more subjective than objective data. Examples range from pricing data, information from customers about the product or service offered by a company, and concrete information about management compensation, financial performance, and similar data. The metaphor of a wave is compelling but data within would be helpful.
Second, the notion of “real data” underscores that talk about data is often just that—chatter, jargon, baloney. “Real data” are difficult to obtain. For example, a company provides a system which tracks and indexes content in the “hidden Web.” What’s the benchmark? How much data are tracked? How much are not indexable? Other questions like this can be answered but time and money are one hurdle. The real reason is that no one wants to make the effort to get data which can be analyzed and then evaluated in head to head comparisons. “Real data”, such as information spewed from financial analysis spreadsheets, is not examined with care. Dig in and the numbers can wobble. Did a scrutinized company actually cut expenses, or does the spreadsheet report that data in bucket A went away and data in bucket B became larger?
Third, the write up itself emphasizes that visualization, not grubby numbers is where the action is. The future of analysis may be an anigif showing the harried decision maker what he or she needs to know. Who has time to work through data by hand, then comparing those data to other information from other sources?
Quite a write up. Interesting implications. Subjective analysis washes away facts in my experience.
Stephen E Arnold, March 13, 2019
Good News about Big Data and AI: Not Likely
February 25, 2019
I read a write up which was a bit of a downer. The story appeared in Analytics India and was titled “10 Challenges That Data Science Industry Still Faces.” Oh, oh. Maybe not good news?
My first thought was, “Only 10?”
The write up explains that the number one challenge is humans. The idea that smart software would solve these types of problems: Sluggish workers at fast food restaurants, fascinating decisions made by entry level workers in some government bureaus, and the often remarkable statements offered by talking heads on US cable TV “real news” programs, among others.
Nope. The number one challenge is finding humans who can do data science work.
What’s number two after this somewhat thorny problem? The answer is finding the “right data” and then getting a chunk of data one can actually process.
So one and two are what I would call bedrock issues: Expertise and information.
What about the other eight challenges. Here are three of them. I urge you to read the original article for the other five issues.
- Informing people why data science and its related operations are good for you. Is this similar to convincing a three year old that lima beans are just super.
- Storytelling. I think this means, “These data mean…” One hopes the humans (who are in short supply) draw the correct inferences. One hopes.
- Models. This is a shorthand way of saying, “What’s assembled will work.” Hopefully the answer is, “Sure, our models are great.”
Analytics India has taken a risk with their write up. None of the data science acolytes want to hear “bad news.”
Let’s federate and analyze that with great data we can select to generate a useful output. Maybe 80 percent “accuracy” on a good day?
Stephen E Arnold, February 25, 2019
Gartner Does the Gartner Thing: Mystical Augmented Analytics
February 19, 2019
Okay, okay, Gartner is a contender for the title of Crazy Jargon Creator 2019.
I read “Gartner: Augmented Analytics Ready for Prime Time.” Yep, if Datanami says so, it must be true.
Here’s the line up of companies allegedly in this market. I put the companies in alphabetical order with the Gartner objective, really really accurate BCG inspired quadrant “score” after each company’s name. Ready, set, go!
BOARD International—niche player
Birst—niche player
Domo—niche player
GoodData—niche player
IBM—niche player
Information Builders—niche player
Logi Analytics—niche player
Looker—niche player
MicroStrategy—challenger
Microsoft—leader
Oracle—niche player
Pyramid Analytics—niche player
Qlik—leader
SAP—visionary
SAS—visionary
Salesforce—visionary
Sisense—visionary
TIBCO Software—visionary
Tableau—leader
ThoughtSpot—leader
Yellowfin—niche player
Do some of these companies and their characterization—sorry, I meant really really objective inclusion—strike you as peculiar? What about the mixing of big outfits like IBM which has been doing Fancy Dan analytics decades before it acquired i2 Ltd. Analyst’s Notebook? I also find the inclusion of SAS a bit orthogonal with the omission of IBM’s SPSS, but IBM is a niche player.
That’s why Gartner is the jargon leader at this point in 2019, but who knows? Maybe another consulting firm beating the bushes for customers will take the lead. The year is still young.
Stephen E Arnold, February 19, 2019
Analytic Hubs: Now You Know
January 30, 2019
Gartner Group has crafted a new niche. I learned about analytic hubs in Datanami. The idea is that a DMSA or data management solution fro analytics is now a thing. Odd. I thought that companies have been providing data analytics hubs for a number of years. Oh, well, whatever sells.
The DMSA vendor list in “What Gartner Sees in Analytic Hubs” is interesting. Plus the write up includes one of the objective, math based, deeply considered Boston Consulting Group quadrants which make some ideas so darned fascinating. I mean Google. An analytics hub?
Based on information in the write up, here are the vendors who are the movers and shakers in analytic hubs:
Alibaba Cloud
Amazon Web Services
Arm
Cloudera
GBase
Hortonworks
Huawei
IBM
MapR Technologies
MarkLogic
Micro Focus
Microsoft
Neo4
Oracle
Pivotal
SAP
Snowflake
Teradata
This is an interesting list. It seems the “consultants” at Gartner, had lunch, and generated a list with names big and small, known and unknown.
I noted the presence of Amazon which is reasonable. I was surprised that the reference to Oracle did not include its stake in a vendor which actually delivers the “hubby” functions to which the write up alludes. The inclusion of MarkLogic was interesting because that company is a search system, an XML database, and annoyance to Oracle. IBM is fascinating, but which “analytic hub” technology is Gartner considering unknown to me. One has to admire the inclusion of Snowflake and MapR Technologies.
I suppose the analysis will fuel a conference, briefings, and consulting revenue.
Will the list clarify the notion of an analytics hub?
Yeah, that’s another issue. It’s Snowflake without the snow.
Stephen E Arnold, January 30, 2019
Big Data Answers the Question ‘Are You Happy?’
November 30, 2018
navigate to the capitalist tool and read “Mapping World Happiness 2015-2018 Through 850 Million News Articles.” Keep in mind that the write up does not explain what percentage of the “news articles” are fake news, the outputs of anti American interest groups, bots, public relations outfits like Definers, or marketing wizards chugging along in the search engine optimization game, and other interesting sources of the data. The write up is a bit of promotion for what is called the GDelt Project. The exercise reveals some of the strengths of open source intelligence. The idea is that collection and analysis can reveal trends and insights.The process involved some number crunching; for example:
Its sentiment mining system has computed more than 2.3 trillion emotional assessments across thousands of distinct emotions from “abashment” to “wrath.”
Google apparently contributed resources.
The question becomes, “Is this analysis an example of real news or is it more marketing?”
The Beyond Search goose has no strong “sentiment” either way. Just asking a simple question.
Stephen E Arnold, November 30, 2018
Making Sense of Big Data: What Is Needed Now
October 29, 2018
Picture, images, and visualization will chop Big Data down to size. SaveDelete explained this idea in depth in its recent story: “The Next Big Phase of Big Data: Simplification.”
According to the article:
“Data visualization is a growing trend, and that momentum isn’t likely to decline anytime soon. Visuals make everything simpler; complex relationships between data points can be seen at a glance, reporting is reduced to a handful of pages, and the esoteric mathematics and statistics behind variable relationships disappear when you’re communicating with someone inexperienced.”
Other ways to deal with making sense of Big Data include:
- “Approachable” software. I think this means easy to use, maybe?
- Gathering the right data. Yep, if one wants to understand terrorist attacks one does not need too much data about hamburger sales in downtown Louisville.
- Reducing insights. This is a tough one. I think the idea is similar to Admiral Craig Hosmer’s statement to me in 1973: “If you can’t get it on a 4×6 note card, I don’t want to see it.”
- Make everything simple. Homer Simpson would be proud.
Useful for math and statistics majors.
Stephen E Arnold, October 29, 2018
Cognos Gets a Rework
October 25, 2018
Cognos? Cognos?
Oh, right, that’s the Canadian analytics company founded in 1969. I think that works out to 49 years young. IBM has owned Cognos since 2008, Now after a decade of vast investment, savvy upgrades, and stellar management decisions, Cognos is going to get even better. Think of it as a US professional football player from the 1960s, suiting up and starting for the Kansas City Chiefs or the Chicago Bears. That’s a strategy that the opposing teams will find surprising.
Same with advanced analytics. Quid, Palantir, Recorded Future! Are you nervous about the new and improved Cognos revealed in “IBM Integrates Business Intelligence and Data Science with New Major Update to Cognos Analytics.”
What’s the fountain of youth?
According to the write up:
… Storytelling… allows users to create interactive narratives by assembling visualizations into a sequence and then enhancing it with media, web pages, images, shapes, and test.
And:
Smart exploration will help users be able to better understand what’s behind their results by analyzing it with machine learning and pattern detection.d then enhancing it with media, web pages, images, shapes, and test.
And:
advanced analytics that include predictive analytics, the ability to identify data patterns and variables driving a certain outcome, smart annotation, and natural language generated insights of data.
But the number one enhancement is… wait for it….
The key new features of this release are a new AI Assistant and pattern detection capability. The AI Assistant enables users to make queries and then receive results in natural language. According to IBM, this makes it easier to not only look for answers, but understand where they come in.
Ah, IBM. Making a product that is half a century young even more appealing to millennials.
Stephen E Arnold, October 25, 2018
Analytics: From Predictions to Prescriptions
October 19, 2018
I read an interesting essay originating at SAP. The article’s title: “The Path from Predictive to Prescriptive Analytics.” The idea is that outputs from a system can be used to understand data. Outputs can also be used to make “predictions”; that is, guesses or bets on likely outcomes in the future. Prescriptive analytics means that the systems tell or wire actions into an output. Now the output can be read by a human, but I think the key use case will be taking the prescriptive outputs and feeding them into other software systems. In short, the system decides and does. No humans really need be involved.
The write up states:
There is a natural progression towards advanced analytics – it is a journey that does not have to be on separate deployments. In fact, it is enhanced by having it on the same deployment, and embedding it in a platform that brings together data visualization, planning, insight, and steering/oversight functions.
What is the optimal way to manage systems which are dictating actions or just automatically taking actions?
The answer is, quite surprisingly, a bit of MBA consultantese: Governance.
The most obvious challenge with regards to prescriptive analytics is governance.
Several observations:
- Governance is unlikely to provide the controls which prescriptive systems warrant. Evidence is that “governance” in some high technology outfits is in short supply.
- Enhanced automation will pull prescriptive analytics into wide use. The reasons are one you have heard before: Better, faster, cheaper.
- Outfits like the Google and In-Q-Tel funded Recorded Future and DarkTrace may have to prepare for new competition; for example, firms which specialize in prescription, not prediction.
To sum up, interesting write up. perhaps SAP will be the go to player in plugging prescriptive functions into their software systems?
Stephen E Arnold, October 19, 2018
Google: Online to Brick and Mortar Cross Correlation
August 31, 2018
Our research suggests that Amazon may have a slight edge in the cross correlation of user data. Google, whether pulling a me too or simply going its own way, has decided to link online and brick and mortar data.
The effort was revealed in “Google and MasterCard Cut a Secret Ad Deal to Track Retail Sales.” Amazon has access to some data which makes it possible for those with appropriate system access to perform analyses of Amazon customers’ buying behavior.
According to the write up:
For the past year, select Google advertisers have had access to a potent new tool to track whether the ads they ran online led to a sale at a physical store in the U.S. That insight came thanks in part to a stockpile of MasterCard transactions that Google paid for. But most of the two billion MasterCard holders aren’t aware of this behind-the-scenes tracking. That’s because the companies never told the public about the arrangement.
To be fair, I am not sure any of the financial services and broker dealer firms provide much output about the data in their possession, who has access to these data, and what use cases are applicable to these data.
From my vantage point in Harrod’s Creek, Kentucky, Google can find its own use cases for Mastercard data.
One question: Does Mastercard pay Amazon to process its data, or does Amazon pay Mastercard?
Google, if the information in the real news article is accurate, is paying for data.
I will address Amazon’s real time streaming data marketplace in my upcoming lecture in Washington, DC. If the information in the US government document I cite in my talk in correct, Google has to shift into high gear with regard to cross correlation of shopper data.
Stephen E Arnold, August 31, 2018
-
- Subscribe to Beyond Search
Feature archive
News archive
-
