Archive for the ‘Big data’ Category

Happy? Sad? Excited? … Facebook can tell.

May 16, 2017

And, has been caught doing just that.

=======

It always amazes me what people post on Facebook. Their daily activities, their deepest emotions – you name it.

By now, every Facebook user should know that FB sifts through their content – posts, pictures, links, emojis – to determine, for example, what topics are hot; what people are doing; which brands people are buying, recommending, trashing or considering; whether users are feeling happy, sad, scared, excited.

The latter is called “sentiment analysis” using computer algorithms to take users’ “emotional pulse”.

Of course, FB promises that they’ll protect users’ privacy and would never even consider divulging that information to outsiders, say, advertisers or political campaigns.

image

Bad news for believers: FB was caught “sharing” sentiment analysis data.

=========

According to USA Today

Documents leaked to a newspaper, The Australian, indicate that Facebook executives prepared a report for one of the country’s top banks.

The report described how Facebook gleans psychological insights into the mood shifts of millions of young people in Australia and New Zealand by monitoring their status updates and photos.

The 23-page report showed Facebook’s ability to detect when users as young as 14 are feeling emotions such as defeat, stress, anxiety or being overwhelmed … and. other information on young people’s emotional well-being such as when they exhibit “nervous-excitement” are “conquering fears“.

FB claimed that it can track how emotions fluctuate during the week.

Anticipatory emotions are more likely to be expressed early in the week.

Reflective emotions increase on the weekend.

Monday-Thursday is about building confidence.

The weekend is for broadcasting achievements.

At a relatively benign level, advertisers can use that information to target ads to certain age groups … and they can time them to run on a certain day.

That’s apparently what FB got caught doing – revealing anonymous and aggregated data – to a potential advertising client.

========

Let’s go a step further…

According to the article: “Facebook has also come under heavy scrutiny in the past for secretly conducting research that manipulated the emotions of users by altering what they see in their News Feed without their consent.”

So, it doesn’t take much creativity to imagine the collection and dissemination of individuals’ sentiment data that could be used to target advertising to specific individuals at specific times – say, when they’re feeling down and are vulnerable to buying certain products geared to giving them a pick-me-up, say, some new clothes, a fancy car or miracle drug.

Pretty unnerving, right?

Of course, FB assures users that it would never consider divulging that sort of data.

Yeah, right.

========

Connecting dots

In a prior post, we reported on a study that concluded time on Facebook can be hazardous to your mental health.

For details see Studies: More time on Facebook … and it’s not good for you.

So, being on Facebook can make you emotionally vulnerable.

Facebook can determine when you’re vulnerable.

Facebook can sell that info to advertisers.

But, FB assures us that it won’t sell that data.

Whew … that’s a relief.

========

#HomaFiles

Follow on Twitter @KenHoma            >> Latest Posts

========

The limits of data analytics …

April 26, 2017

Team Clinton worshiped at the altar and got burned.

======

One of the themes in the book Shattered was that theClinton campaign got fixated on their data-rich electorate models, applying the models robot-like to allocate ad dollars, deploy field workers and schedule “market visits” by Hillary and her surrogates.

clip_image002

What Team Clinton seemed to have forgotten is the old Reagan adage: trust but verify.

The data models – which worked near-flawlessly for Obama – took the stage as “shiny objects” that led the Clinton campaign astray.

Politico reported a case study that  illustrated the point …

(more…)

Nums: A world of battling algorithms

February 21, 2017

In my SBA course, we explored how human judgment and decision-making can often be outperformed by out-performed by algorithms, especially in oft-repeated data-rich situations which are largely rules-based.

In a cool 15 minute TED Talk (my all time favorite), tech entrepreneur Kevin Slavin tells how algorithms have reached across industries and into every day life.

A couple of lines caught my attention:

  • There are more than 2,000 physicists working on Wall Street developing operational algorithms
  • Massive scale speed trading is dependent on millisecond read & respond rates …
  • So, firms are physically literally locating right next to internet routing hubs to cut transmission times
  • And, of course, there isn’t time for human intervention and control
  • “We may be building whole worlds we don’t really understand, and can’t control.”

Worth listening to this pitch … a very engaging geek who may be onto something big.

click  to view video
image

* * * * *
Follow on Twitter @KenHoma        >> Latest Posts

The limits of data analytics …

January 19, 2017

Team Clinton worshipped at the altar and got burned.

======

Lots of post-election articles about how the Clinton campaign got fixated on their data-rich electorate models, using them to allocate ad dollars, deploy field workers and schedule “market visits” by Hillary and her surrogates.

clip_image002

What Team Clinton seemed to have forgotten is the old Reagan adage: trust but verify.

The data models – which worked near-flawlessly for Obama – took stage as “shiny objects” that led the Clinton campaign astray.

Politico reported a case study that  illustrated the point …

(more…)

Flashback: Remember when Target caused a stir by ID’ing moms-to-be?

June 15, 2016

Now, researchers are trolling your web searches to auto-detect diseases.

The Washington Post recently channeled a study done by Microsoft — published in the Journal of Oncology Practice …

The essence: Microsoft’s big data analysts ID’ed folks who were querying questions like “how to treat pancreatic cancer” — hypothesizing that they might have been diagnosed with the disease.

Then, the researchers backtracked thru the prior searches done by those folks and detected a pattern of precedent queries that revolved around symptoms, e.g. abdominal swelling.

Bottom line:  the researchers were able to use the inferred pattern of symptoms to early-predict a disease diagnosis for a statistically significant number people who queried symptoms.

That’s potentially big news in disease diagnosis, though doctors caution that for many diseases, the onset of patient-queried symptoms may be too late-stage for effective treatment.

 

image

======
The Microsoft query- disease analysis reminded me of how Target created some Big Data buzz for analyzing purchase patterns to ID moms-to-be. 

(more…)

Hacked: Cards expose Moneyball’s strategic vulnerabilities …

June 18, 2015

Moneyball – the Oakland As use of data  and metrics to ID undervalued players —  was one of the  major catalysts for the current rage around big data and data analytics.

The Houston Astro’s  were one of the teams to adopt the Moneyball philosophy in a big way.

This week, the NY Times broke the story that the St. Louis Cardinals had hacked into Astro’s proprietary database.

Big news.

In fact, this hack seemed to get more media time than  the Chinese jacking the personal info of all government employees.

Hmmm.

image

Baseball competition aside, here’s why I think there’s a big teaching point in the story

(more…)

Mouse tracks: Mickey’s hot on your trail …

February 20, 2015

According to Business Week

“Disney has launched a $1 billion experiment in crowd control, data collection, and wearable technology that could change the way people play—and spend—at the Most Magical Place on Earth. “

 

image

The innovation – called MyMagic – let’s Mickey track every move you make around the old Magic Kingdom.

(more…)

Mouse tracks: Mickey’s hot on your trail …

April 14, 2014

According to Business Week

“Disney has launched a $1 billion experiment in crowd control, data collection, and wearable technology that could change the way people play—and spend—at the Most Magical Place on Earth. “

 

image

The innovation – called MyMagic – let’s Mickey track every move you make around the old Magic Kingdom.

(more…)

Nums: Ask why … not just how many.

April 10, 2014

Some highlights from an HBR article:  The Hidden Biases in Big Data

These days the business and management science worlds are focused on how large datasets can decode consumers’ behavior patterns … enabling marketers to laser-target high potential prospects with finely-honed messages, offers, and “attention”.

“Big data” … becomes problematic when it adheres to “data fundamentalism” … the notion that correlation always indicates causation, and that massive data sets and predictive analytics always reflect objective truth … that  “with enough data, the numbers speak for themselves.”

image

Big data has hidden biases in both collection methods and analysis …

(more…)

Nums: Ask why … not just how many.

February 7, 2014

Some highlights from an HBR article:  The Hidden Biases in Big Data

These days the business and management science worlds are focused on how large datasets can decode consumers’ behavior patterns … enabling marketers to laser-target high potential prospects with finely-honed messages, offers, and “attention”.

“Big data” … becomes problematic when it adheres to “data fundamentalism” … the notion that correlation always indicates causation, and that massive data sets and predictive analytics always reflect objective truth … that  “with enough data, the numbers speak for themselves.”

image

Big data has hidden biases in both collection methods and analysis …

(more…)

Flashback: Remember when Target caused a stir by aiming at moms-to-be?

June 11, 2013

The Feds’ phone & internet surveillance programs that were revealed last week have raised the public’s consciousness re: Big Data.

Remember when Target created some Big Data buzz for analyzing purchase patterns to ID moms-to-be?

image

In a previous post, we excepted from a NY Times article How Companies Learn Your Secrets that

  1. Much of what people do is based on habits, not conscious reasoning.
  2. Consumers’ shopping habits and brand loyalties are often more habitual than thoughtful.
  3. But, there are certain “events” — e.g. new baby, new home, recent divorce — that seem to make consumers more open to switching stores and brands.
  4. Savvy marketers are learning to identify these critical events — before they happen — and try to get consumers to switch  their behavior.

Target is one of the retailers identifying customers who are “vulnerable to intervention by marketers” … and pouncing on them.

Who?  Moms-to-be.

How are they doing it?

(more…)

Nums: Ask why … not just how many.

April 10, 2013

Some highlights from an HBR article:  The Hidden Biases in Big Data 

These days the business and management science worlds are focused on how large datasets can decode consumers’ behavior patterns … enabling marketers to laser-target high potential prospects with finely-honed messages, offers, and “attention”.

“Big data” … becomes problematic when it adheres to “data fundamentalism” … the notion that correlation always indicates causation, and that massive data sets and predictive analytics always reflect objective truth … that  “with enough data, the numbers speak for themselves.”

image

Big data has hidden biases in both collection methods and analysis …

(more…)

What does this map represent?

January 7, 2013

Take a guess …

image

No good reason for anybody to know.

It’s a mapgraphic depicting the 1,540 Walmart stores in 1990.

So what?

Here’s what makes it interesting.

For a cool, dynamic visual showing how & where Walmart has grown over the years, click the link to view FlowData.com’s Walmart growth map.

The content is interesting, and it’s a nice way to present geo-time series data over time.

* * * * *
Follow on Twitter @KenHoma         >> Latest Posts

Nums: A world of battling algorithms

December 27, 2012

In a cool 15 minute TED Talk, tech entrepreneur Kevin Slavin tells how algorithms have reached across industries and into every day life.

A couple of lines caught my attention:

  • There are more than 2,000 physicists working on Wall Street developing operational algorithms
  • Massive scale speed trading is dependent on millisecond read & respond rates …
  • So, firms are physically literally locating right next to internet routing hubs to cut transmission times
  • And, of course, there isn’t time for human intervention and control
  • “We may be building whole worlds we don’t really understand, and can’t control.”

Worth listening to this pitch … a very engaging geek who may be onto something big.

click the pic to view video
image

* * * * *
Follow on Twitter @KenHoma        >> Latest Posts

Behavioral analytics … bad when Target does it … OK for political campaigns?

September 19, 2012

A couple of months ago Target got some bad press when it was revealed that the company was mining customers’ purchase histories to slot them into behavioral groups susceptible to tailored promotional pitches.

For example, Target identified purchases that mothers-to-be made early in their pregnancies – sometimes before they even knew they were pregnant.  Think bigger jeans, skin care lotions.

Many folks railed that it was an example of big brother invasion of privacy.

Well, guess what?

Political campaigns are using the same methods that Target was using

The modern science of politics is increasingly based on principles from behavioral psychology and data analytics.

image

Campaigns today mine large data bases with mathematical algorithms that slot folks into categories and provide the basis for how people should be approached (or ignored).

According to the WSJ:

Perhaps the most valuable data in modern campaigns comes from statistical “microtargeting” models—the political world’s version of credit scores.

Campaigns gather thousands of data points on voters, culled from what they put on their registration forms, what they have told canvassers who have visited their homes in the past, and information on their buying and lifestyle habits collected by commercial data warehouses.

The campaigns then run algorithms trawling for patterns linking those demographic characteristics to the political attitudes measured in their polling.

Financial institutions run such statistical models to generate predictions about whether a given individual will demonstrate a certain behavior, like paying a bill on time or defaulting on a loan.

Campaigns do the same, only they are interested in predicting political behavior.

So it’s typical now to generate individual scores, presented as a percentage likelihood, that a voter will cast a ballot, support one party or the other, be pro-choice or antiabortion, or respond to a request to volunteer.

These scores now stick to voters as indelibly as credit scores.

And just as a bank officer won’t sign off on a loan without requesting one, a field director for a campaign won’t send a volunteer to a voter’s door without knowing the relevant number.

BTW: It’s Team Obama that’s doing most of this stuff.

Bad for Target … but OK for Obama.

Hmmm

* * * * *

WSJ source: “The Victory Lab: The Secret Science of Winning Campaigns” by Sasha Issenberg

>> Latest Posts