March 22nd 2017

Book Review: Weapons of Math Destruction by Cathy O’Neil

Cathy ONeal 3 17 2017

 

By: Eric Siegel, Founder, Predictive Analytics World

Originally published in Analytics Magazine

Book: Weapons of Math Destruction: How Big Data Increases Inequality and Threatens Democracy by Cathy O'Neil (Crown, September 2016)

Book review bottom line: Definitely go read this book, despite the fact that it does convey a certain oversimplifying, "black-and-white" position.

Related reading: I devoted all of Chapter two of Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die (2013; 2016) to ethical issues that arise in predictive analytics’ deployment. Parts of that chapter, along with my article “The Risk of Prejudice in Computerized Prediction” in Profiles in Diversity Journal, cover some of the same topics of concern in Weapons of Math Destruction.

My Review of This Book:

Cathy O'Neil's New York Times Bestseller Weapons of Math Destruction belongs squarely in the "must read" category. In this first-of-its-kind book, the author, an industry insider and experienced expert, thoroughly covers the sociological downside of data science.

In the world of big data, there's a lot of music to be faced. With all its upside, data science's deployment risks being prejudicial, predatory, exploitative, buggy, blindly trusted, and secretive. And it has the potential to magnify the consumer’s personal economic struggle rather than remedy it.

These risks permeate across the field. The book's broad coverage includes examples from all the main business application areas to which predictive models commonly apply: marketing, online ads, credit scoring, insurance, workforce analytics, law enforcement, and political campaigns.

By providing such a uniquely comprehensive treatment of data's downside, this book addresses two dire needs: increasing awareness and opening the door to prolific discussion.

Click here to access Eric Siegel’s full book review in Analytics Magazine (above are only the first four of the review’s 21 paragraphs).

No Comments yet »

March 20th 2017

Wise Practitioner – Predictive Analytics Interview Series: Angel Evan at Angel Evan, Inc.

By: Eric Siegel, Founder, Predictive Analytics World

In anticipation of his upcoming conference co-presentation, Identifying Unique Gamer Angel Evan PAW BLOGTypes Using Predictive Analytics at Predictive Analytics World San Francisco, May 14-18, 2017, we asked Angel Evan, Founder of Angel Evan, Inc., a few questions about his work in predictive analytics.

Q: In your work with predictive analytics, what behavior or outcome do your models predict?

A: I work in marketing, so typically we are trying to predict a customer behavior or outcome. For example, whether or not a customer is likely to cancel a subscription or the probability they will respond to an ad. 

Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?

A: If I were to cite just one thing, I would say that it helps us (and our clients) figure out where to focus, i.e., which customers are most at risk of leaving, or which customers represent the most potential revenue. 

Q: Can you describe a quantitative result, such as the predictive lift of your model or the ROI of an analytics initiative?

A: While I cannot share actual figures due to NDA, I can say that we recently finished creating a predictive model for a large wine brand to determine which customers were most likely to cancel their wine club memberships. Wine clubs are a vital revenue stream for most wineries, as they represent the single highest percentage of profitability. As a result of our analysis and predictive model, the company saw a decrease in customer churn in just 60 days. 

Q: What surprising discovery or insight have you unearthed in your data?

A: Which predictive variables cause individual customer segments to act the way that they do.

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A: The old data attributes of media planning, especially household income, don’t necessarily influence customer spending the way people think. Traditionally media and strategy planners would target people with high household incomes based on the perception that they have more disposable income. What we’re seeing is that user behavior is a better predictor of outcomes. 

———————

Don't miss Angel’s conference co-presentation, Identifying Unique Gamer Types Using Predictive Analytics on Tuesday, May 16, 2017 at 11:45 am to 12:05 pm at Predictive Analytics World San Francisco. Click here to register to attend

By: Eric Siegel, Founder, Predictive Analytics World

No Comments yet »

March 17th 2017

Wise Practitioner – Predictive Analytics Interview Series: Paul Speaker at The Dow Chemical Company

By: Eric Siegel, Founder, Predictive Analytics World

In anticipation of his upcoming conference presentation, Creating an industrial Revolution for Analytics at Predictive Analytics World San Francisco, May 14-18, 2017, we asked Paul Speaker, Senior Data Scientist at The Dow Chemical Company, a few questions about his work in predictive analytics.

Q: In your work with predictive analytics, what behavior or outcome do your models predict?

A: Our group’s work is involved in developing predictive and prescriptive analytics in the sales, commercial, and supply chain spaces. 

Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?

A: We have a set of predictive models, which, when used together, can do anything from drive demand planning to prioritize the targeting of customers to prevent churn or other leakage.

Q: What surprising discovery or insight have you unearthed in your data?

A:  Looking at data at the transactional level always reveals behavior not just of customers but of everyone who is involved in “creating” the data.

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A: While the primary focus of innovation in analytics has been from increasingly complex predictive algorithms, the creative combination of predictive results of relatively simple models can result in even more value.

———————

Don't miss Paul’s conference presentation, Creating an industrial Revolution for Analytics on Tuesday, May 16, 2017 from 3:55 to 4:15 pm at Predictive Analytics World San Francisco. Click here to register to attend.

By: Eric Siegel, Founder, Predictive Analytics World

No Comments yet »

March 10th 2017

Wise Practitioner – Predictive Analytics Interview Series: George Iordanescu at Microsoft

By: Bala Deshpande, Co-Chair, Predictive Analytics World Manufacturing

In anticipation of his upcoming conference presentation, Predictive Analytics Solution George Iordanescu Blog PageTemplate for Early Prediction of Assembly Line Failures, at Predictive Analytics World Manufacturing Chicago, June 19-22, 2017, we interviewed George Iordanescu, Data Scientist at Microsoft.  View the Q-and-A below for a glimpse of what’s in store at the PAW Manufacturing conference.

Q: What are the challenges in translating the lessons of predictive analytics from other verticals into manufacturing?

A: Rather than challenges, from the Data Science perspective I would say there are actually exciting opportunities. The Cortana Intelligence solution How-to Guide I am presenting showcases an important aspect that is specific to assembly line failures data: We provide a generic Advanced Analytics solution that uses Machine Learning to predict failures before they happen. Early prediction of future failures allows for less expensive repairs or even discarding, which are usually more cost efficient than going through recall and warranty cost.

Q: In your work with predictive analytics, what behavior do your models predict?

A: A key point is that customers’ pain points – the failures- and their subject matter expertise are actually the real goldmine. By specifically looking at returns and functional failures at the end of assembly line and combining these with domain knowledge and root cause analysis we provide a generic advanced analytics solution with a modular design that encapsulates main processing steps and leverages test systems already in place. So no new equipment is necessary, since we use machine learning to detect subtle patterns in test systems measurements that pass the regular quality checks but are still indicative of a future failure.

Q: How does predictive analytics deliver value at your organization? What is one specific way in which it actively drives decisions?

A: A good scenario is the OEMs who use contract manufacturers (Apple-Foxconn, MSFT-Jabil), and want to minimize post sale service and warranty costs: Use ML to build models that use test/shop floor data (that belongs to OEM and or CMs) to predict before shipping, field return and repairs that may happen months or years after the device is shipped. This enables predicting future failures while the device is still in the early manufacturing line stages or is already assembled but is not yet shipped, so that fixing or even discarding it may be cheaper than going through recall and warranty costs.

Q: Can you describe a successful result, such as the predictive lift (or accuracy) of your model or the ROI of an analytics initiative?

A: A good example is Jabil where they are not only predicting that a potential issue or failure could occur, but by having visual insight into all levels of production and operations, the company also now knows why the failure was predicted. That it is actionable information Jabil can use to avoid a costly loss, while shortening product lead times and delivering superior quality.

Q: What surprising discovery have you unearthed in your data?

A: I think it is surprising and very empowering to see that fundamental information comes from the client: Failure data, domain knowledge, and test systems already in place. This setup happens to be a natural example of a Machine Learning concept called Gradient Boosting, where an imperfect (weak) model (i.e., the existing quality check system) is extended and improved by focusing on its errors, i.e., the hard cases given by the assembly line failures.

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World for Manufacturing.

A: We show how a solution can be implemented and deployed fast in the cloud, using the flexible on-line Microsoft Azure platform that decouples infrastructure components (data ingestion, storage, data movement, visualization) from analytics engine that supports modern data science languages like R and Python. The solution modeling component can thus be retrained as needed and be implemented using high performance Azure Machine Learning algorithms, or open source (R/Python) libraries, or from a third-party solution vendor.

———————

Don't miss George’s conference presentation, Predictive Analytics Solution Template for Early Prediction of Assembly Line Failures on Tuesday, June 2017 at 10:30 to 11:15 am at Predictive Analytics World Manufacturing Chicago. Click here to register to attend.

By: Bala Deshpande, Founder, Simafore and Conference Co-Chair of Predictive Analytics World Manufacturing

No Comments yet »

March 3rd 2017

Wise Practitioner – Predictive Analytics Interview Series: Afsheen Alam at Allstate Insurance

By: Eric Siegel, Founder, Predictive Analytics World

In anticipation of her upcoming conference presentation, Our Success with Agile Afsheen Alam IMAGE 3Analytics at Predictive Analytics World for Business Chicago, June 19-22, 2017, we asked Afsheen Alam, Program Manager Marketing Analytics and Big Data at Allstate Insurance, a few questions about her work in predictive analytics.

Q: In your work with predictive analytics, what behavior or outcome do your models predict?

A: Insurance quoting and binding behavior.

Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?

A: Lead management, Direct Mail marketing, Digital Marketing channels.

Q: What surprising discovery or insight have you unearthed in your data?

A: We have several models with various lines of business and some variables keep showing repeatedly across all lines of business and various models.

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A: My talk is about Agile Applications to the Modeling process. It will explain how a streamlined process can deliver a pipeline of models in a relatively short period of time by having structured sprints with defined deliverables. A model is only as good as its speed to market and execution for business results. The Agile process helps deliver those results in a creative environment.

———————

Don't miss Afsheen’s conference presentation, Our Success with Agile Analytics on Tuesday, June 20, 2017 from 3:55 to 4:15 pm at Predictive Analytics World Chicago. Click here to register to attend.

By: Eric Siegel, Founder, Predictive Analytics World

No Comments yet »

February 27th 2017

Wise Practitioner – Predictive Analytics Interview Series: Jennifer Bertero at CA Technologies

By: Eric Siegel, Founder, Predictive Analytics World

In anticipation of her upcoming conference presentation, Redefining Analytics for Jennifer Bertero IMAGE 2Marketing, at Predictive Analytics World San Francisco, May 14-18, 2017, we asked Jennifer Bertero, VP, Business Analytics at CA Technologies, a few questions about her work in predictive analytics.

Q: In your work with predictive analytics, what behavior or outcome do your models predict?

A: Our analytics work has been focused mostly on the magic triangle of sales, marketing and customer success. We’ve been working on helping to improve customer experience with the customer team. We’ve helped to prioritize sales efforts on opportunities and accounts that are most likely to buy. We’re also helping marketing to identify what prospects are ready for the sales process and how they can more effectively nurture leads.

Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?

A: Analytics is pervasive in decision making. We’re influencing strategic, as well as operational decisions. For example, analytics to predict which accounts have high propensity to buy our products have been used by sales to devise a customer segmentation strategy. Analytics that predict whether a sales opportunity will be won or lost is used by sales to prioritize which opportunities they should target. Customer experience analytics is helping our product teams design their products with customer experience in mind. In addition, it’s helping identify those customers early who are not likely to renew their contracts.

Q: Can you describe a quantitative result, such as the predictive lift of your model or the ROI of an analytics initiative?

A: This is one of the most difficult but important parts of the analytics process. We’ve learned to identify up front what metrics we’re going to track and what success looks like – where we can. There are also discovery projects that yield insights that then lead to business process change. For example, the insights we provided to the customer engagement team allowed us to track intent, activity, and questions answered in our online communities. We correlated these to the expensive calls coming into the support line. We helped influence the business to increase support engineer engagement in communities to solve problems before they ever reach for the phone. We increased questions answered by 28%, helped decrease support call costs and increase by double digits our overall company NPS score.

Q: What surprising discovery or insight have you unearthed in your data?

A: We discovered that if a customer's question about our products is not answered in the digital communities within 2 days, our support desk received a call with the same question. This led to the business process change and improvements I just described in customer support. In marketing, we discovered that most of our leads were coming in from small to medium-sized businesses but our marketing efforts and sales process was built more for large customers. We used those insights to tailor our campaigns more effectively and funnel leads to the right sales teams by the right products.

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A: We’ve built our analytics shop from ground up and now can hardly keep up with the demand from the business. After consistently delivering value for two years, we’re now being asked to help shape the strategy of the company, using insights and analytics to help the C level make better and data driven business decisions. Coming from the business side, I understand the business of trying to compete and win in an extremely competitive, fast-paced industry. I’ll share the key business metrics to focus on when working on analytics projects and how they can directly help the business strategy and performance. Analytics can be used as a secret weapon. Today, it’s not a nice-to-have; it’s a must-have. It’s nice to be needed but it’s more important to be relevant, timely and directly tied to success.

———————

Don't miss Jennifer’s conference presentation, Redefining Analytics for Marketing on Wednesday, May 17, 2017 from 11:15 am to 12:00 pm at Predictive Analytics World San Francisco. Click here to register to attend.

By: Eric Siegel, Founder, Predictive Analytics World

No Comments yet »

February 23rd 2017

Wise Practitioner – Predictive Analytics Interview Series: Michael Dessauer at The Dow Chemical Company

By: Eric Siegel, Founder, Predictive Analytics World

In anticipation of his upcoming conference presentation, Listening Down the Value Chain: Michael Dessauer IMAGE 2Using Text-based Predictive Models to Find New Opportunities for B-to-B Business, at Predictive Analytics World San Francisco, May 14-18, 2017, we asked Michael Dessauer, Data Scientist at The Dow Chemical Company, a few questions about his work in predictive analytics.

Q: In your work with predictive analytics, what behavior or outcome do your models predict?

A: Our market listening modeling efforts focus on two types of predictions: consumer sentiment relative to our current value propositions and consumer needs to validate / identify future technology focus areas.

Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?

A: One specific example is identifying cross-selling opportunities using our custom-developed recommender engine.

Q: Can you describe a quantitative result, such as the predictive lift of your model or the ROI of an analytics initiative?

A: We have assisted businesses on price optimization which significantly lifted their margins while reducing business resources needs.

Q: What surprising discovery or insight have you unearthed in your data?

A: We have been able to better quantify a market disruption by eCommerce-exclusive brands which was surprising in its significance.

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A: The audience will hopefully leave with a clear understanding of market listening’s importance and value for B-to-B businesses through real-life examples.

———————

Don't miss Michael’s conference presentation, Listening Down the Value Chain: Using Text-based Predictive Models to Find New Opportunities for B-to-B Business on Tuesday, May 16, 2017 4:20 to 4:40 pm, at Predictive Analytics World San Francisco. Click here to register to attend.

By: Eric Siegel, Founder, Predictive Analytics World

No Comments yet »

February 17th 2017

Wise Practitioner – Predictive Analytics Interview Series: Steven Ulinski at Health Care Service Corporation

By: Eric Siegel, Founder, Predictive Analytics World

In anticipation of his upcoming conference presentation, Challenges of Information and Steve Ulinski IMAGE 2Cyber Security Using Predictive Analytics at Predictive Analytics World for Business Chicago, June 19-22, 2017, we asked Steven Ulinski, Security Data Scientist at Health Care Service Corporation, a few questions about his work in predictive analytics.

Q: In your work with predictive analytics, what behavior or outcome do your models predict?

A: Honestly, we don’t know yet. We are just beginning to think data driven and move past rule based systems. Ultimately we are looking to predictive analytics to help us identify an attack on our systems before it happens so we can implement changes in our security posture to prevent any data loss. At this time, it’s more about finding breaches faster, and the ability to respond faster.

Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?

A: In the cyber security industry, predicative analysis has provided a new method to identify a security incident, for instance using random forests against historical data. This is still a descriptive analysis, and we are looking for systems and models that will move us past this, and actually identify threats before they become threats. We want to actually move past predictive and get to prescriptive.

Q: Can you describe a quantitative result, such as the predictive lift of your model or the ROI of an analytics initiative?

A: We’re interested in preventing data breaches. We know that predictive analytics will increase our false positives rates. Ultimately, we will rate the value and quality of the models if they can discover data breaches in a timely manner.

Q: What surprising discovery or insight have you unearthed in your data?

A: In the past year we have researched how predictive analysis, other data analysis, and AI systems can help us. We haven’t truly begun to analyze our data with predictive analytics yet. However, I think the biggest challenge that we have with the data is the volume, variety, and velocity. Consider an enterprise firewall infrastructure. They generate a massive amount of information per hour. We want to analyze the firewall information with operating system information, user web behavior, and other factors to have a holistic view for threat detection. We have a lot of information, from various different logging mechanisms. Our analysts just cannot keep up with the data. We are looking to predictive analytics and other systems to help wrangle in our data. This is one reason Information Security systems are leveraging Big Data and ML.

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A: The Rexer Data Science Survey showed that Information Security analytic goals were raised from 3% in 2013 to 5% in 2015. Additionally, only 1% of the time is being dedicated to Information Security. There are significant challenges in analyzing Information Security data. I’ll be discussing more about these challenges.

———————

Don't miss Steven’s conference presentation, Challenges of Information and Cyber Security Using Predictive Analytics on Wednesday, June 21, 2017 from 3:30 to 4:14 pm at Predictive Analytics World Chicago. Click here to register to attend.

By: Eric Siegel, Founder, Predictive Analytics World

No Comments yet »

February 13th 2017

Wise Practitioner – Predictive Analytics Interview Series: Lauren Haynes at The University of Chicago

By: Eric Siegel, Founder, Predictive Analytics World

In anticipation of her upcoming conference presentation, Data Science for Social Good: Lauren Haynes IMAGE 2How Predictive Analytics Can Help Governments and Non-Profits, at Predictive Analytics World for Business Chicago, June 19-22, 2017, we asked Lauren Haynes, Senior Project Manager at Center for Data Science and Public Policy at The University of Chicago, a few questions about her work in predictive analytics.

Q: In your work with predictive analytics, what behavior or outcome do your models predict?

A: We work at the intersection of public policy and predictive analytics – at DSaPP we work in healthcare, social services, governments, non-profits, education, transparency, economic development, public safety, and criminal justice. To that end we help organizations identify the inspections organizations should do to find the highest volume of violations for housing and environmental enforcement, students at risk of dropping out of school, individuals at risk of re-entering the criminal justice system, etc.

Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?

A: We think predictive analytics can help organizations drive decisions of where to use limited resources – if you are responsible for 100,000 buildings and can only inspect 1,000 of them a year, knowing which 1,000 are highest risk for having a violation helps use those inspections effectively. Similarly, if you can only enroll a small population in an intervention or program, being able to identify those most at risk maximizes the value of the intervention. 

Q: Can you describe a quantitative result, such as the predictive lift of your model or the ROI of an analytics initiative?

A: Syracuse deals with hundreds of water main breaks, leaks, and other issues that require attention each year, distributed without a clear pattern across the entire city – in partnership with the city, DSaPP built a model to predict water main breaks. Based on the model’s accuracy at predicting past years’ incidents, the team predicted that 32 of the top 50 highest-risk water mains would break in the next 3 years. If you simply used the age of pipes as a way to prioritize which city blocks should be replaced first, only 5 percent of the top 50 water mains on your list would go on to break in the next 3 years. If you used the history of breaks at different locations, looking at the number of occurrences of breaks per city block, only half of your top 50 riskiest mains would break. But most water main breaks are “first-time offenders,” without prior breaks at that location. Going by past breaks alone, you would never predict any breaks that have previously had less than three water mains breaks, and replacement efforts would only focus on a handful of neighborhoods. In the two weeks after the DSaPP team delivered the risk scores to their Syracuse partners, two of the water mains listed in the top 50 ruptured. The Syracuse Office of Innovation has quickly integrated the model into their work, using it to guide an infrastructure planning process and decide where to do “dig once” combinations of water main replacement and road resurfacing.

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A: Applying data science and predictive analytics to the social sector is as much about change management and organizational readiness as it is about technology.

———————

Don't miss Lauren’s conference presentation, Data Science for Social Good: How Predictive Analytics Can Help Governments and Non-Profits, on Tuesday, June 20, 2017 from 4:45 to 5:30 pm at Predictive Analytics World Chicago. Click here to register to attend.

By: Eric Siegel, Founder, Predictive Analytics World

No Comments yet »

February 10th 2017

Wise Practitioner – Predictive Analytics Interview Series: Daqing Zhao at Macy’s

By: Eric Siegel, Founder, Predictive Analytics World

In anticipation of his upcoming conference presentation, Macy’s Advanced Analytics in Daqing Zhao IMAGE 2Customer Centric Strategies at Predictive Analytics World San Francisco, May 14-18, 2017, we asked Daqing Zhao, Director, Advanced Analytics at Macy’s, a few questions about his work in predictive analytics.

Q: In your work with predictive analytics, what behavior or outcome do your models predict?

A: Macy’s has 150 years of history and is an iconic brand in retail, providing superior omni channel shopping experience for our customers through stores and online channels.  Macy’s makes extensive efforts to protect customer privacy and identity information.  In order to serve our customers better, we need to understand our customer preferences in order to recommend the right product, and send the relevant information, to give our customers the frictionless shopping experience. In particular, we predict what categories, and brand a customer will likely to make a purchase, how many they would spend, as well as how they interact with our marketing channels.  We also predict customer retention, customer life time value and other customer metrics to help our marketing activities.

Q: How does predictive analytics deliver value at your organization – what is one specific way in which it actively drives decisions or operations?

A: Using Predictive Analytics, we are able to best organize our data to predict customer preferences and behaviors, in order to optimize our marketing activities, such as emails, direct mails and product recommendation.

Q: Can you describe a quantitative result, such as the predictive lift of your model or the ROI of an analytics initiative?

A: We do not disclose quantitative numbers.  Predictive analytics, however, is very useful to drive results.  For example, if we predict a customer having high propensities to make a purchase in some categories, and very low propensities to convert on some other categories, we often see differences of an order of magnitude in conversion rates or average spend in these categories.

Q: What surprising discovery or insight have you unearthed in your data?

A: One such surprise is that for some segments of customers, having low scores in all categories may be because we do not have sufficient data about these customers.  It is not necessary that they have no interest in any our categories.  With higher uncertainties in the lower scores, it may not be optimal to make decisions based on the small differences in the scores.

Q: Sneak preview: Please tell us a take-away that you will provide during your talk at Predictive Analytics World.

A: Our customers are diverse and our products and business goals are complex.   We don’t rely on one methodology, one solution, or one platform for our data driven, omni channel, personalized marketing efforts.  We take a portfolio approach in our predictive methodologies, data sources and hypotheses, perspectives and strategies.

———————

Don't miss Daqing’s conference presentation, Macy’s Advanced Analytics in Customer Centric Strategies, on Wednesday, May 17, 2017 from 10:25 to 10:45 am at Predictive Analytics World San Francisco. Click here to register to attend.

By: Eric Siegel, Founder, Predictive Analytics World

No Comments yet »

« Prev - Next »