Team 3: Cody Lundell, Christina Worden, & Daniel Gabiou
The Challenges of Big Data Overview
The human element of the analysis of data collection and application is critically important when considering the challenges
of big data. As explained by Duhigg (2016) through the example of debt collectors with Chase Bank, the data which was
collected and used to increase debt collection among Chase customers was most successful when the debt collectors in a Florida
Chase office actually applied and discussed the meaning of the data to their work. By absorbing the data and also breaking it
down to directly affect their work, they were able to learn about the data and also make it more human in its application.
O’Neil (2016) covers the human element of data collection and use in the introduction of Weapons of Math Destruction. Through
the story of Washington, DC’s Public School Chancellor, Michelle Rhee, O’Neil explains the disconnect of creating a data model
to achieve results within the failing public school system. The model was created to evaluate performance but did not include
proper analysis to fully evaluate the teachers’ performance and capture, for example, instances of bad behavior like teachers
inflating students test scores. This ill-conceived model resulted in teachers who were performing well being fired.
Weapons of Math Destruction: Intro
The Big Data economy is based on math-powered applications that create huge pieces of data from an array of sources and impact
every form of industry. While many view big data as innovative and transformative, O’Neil (2016) cautions that the math
powered applications on which big data relies is traced back to the decisions of fallible humans. O’Neil (2016) gives a name
to these harmful types of models: Weapons of Math Destruction or WMDs for short. Throughout the introduction, O’Neil (2016)
cites instances of big data being misused and as a result causing serious damage. One example she shares is of Michelle Rhee’s
work as Chancellor of Washington, DC’s public school system. Rhee was hired to reform the failing school system and developed
a math-based application to evaluate the teachers to determine who should be fired and who should be rewarded for their
performance. The application was fraught with bad analysis and many teachers who were performing well ended up being fired.
O’Neil (2016) uses this example to reinforce her serious concern about WMDs but also to point out that there is a difference
between math analysis and actual “on-the-ground reality.”
Big Data for Social Innovation
“The term ‘big data’ is used to describe the growing proliferation of data and our increasing ability to make productive use of it.” (Desouza & Smith, 2014, p. 39). The key focus of this article is that data is nothing without knowing how to use it. The authors ask a hypothetical question: could we use big data in a sustainable manner to solve social problems such as homelessness, human trafficking, and education? The four barriers to creating and using big data to solve social problems include: 1) data are buried in administrative systems; 2) data governance standards are lacking; 3) data are often unreliable; and 4) data can cause unintended consequences (Desouza & Smith, 2014, p. 41). Before big data could be utilized to empower evidence-based decision-making for social problems, there needs to be improvements in data collection, organization, and analysis. The authors recommend four solutions to bridge the big data gap: 1) build global data banks on critical issues (such as human trafficking, global hunger, and poverty); 2) engage citizens and citizen science; 3) invest in data curators and analysis training; and 4) promote virtual collaboration (Desouza & Smith, 2014, p. 42-43).
Information is only as good as it can be understood and applied. Information blindness suggests a situation where information is accessible but unable to be applied (Duhigg, 2016, p. 243). In one school district in Ohio, no amount of resources or innovation could help the South Avondale school increase the educational performance of its students. Even after data based on attendance, test scores, homework, and classroom participation was available, nothing changed (Duhigg, 2016, p. 240). After years of unsuccessful initiative, a new solution was offered. The new solution was based on teaching the teachers how to use the data since most admittedly were not even referencing the dashboards. Once the Elementary Initiative (EI) was underway, the results were so positive that the EI became recognized as the model for educational reform. In a similar way, studies have been conducted around why people choose a retirement plan and even how a Chase bank’s debt collection saw great results when they based their behavior on the data.
Key Take-Aways
- Feedback in the form of analysis is vital to ensure that the models being created for big data are properly capturing all data points.
- Data is only useful when it influences behavior.
- Increased emphasis on how to use tools is better than creating more tools.
Discussion Questions
1) For the “Big Data for Social Innovation” reading, do you agree with the authors that improviving data collection, organization, and analysis will help in solving major social problems?
2) In the introduction of Weapons of Math Destruction, O’Neil explains Michelle Rhee’s work as Chancellor of Washington,
DC’s public school system and her failure to reform it using math-based applications. Based on the takeaways from the “Big
Data for Social Innovation” article, what solutions could Rhee have used to bridge the big data gap in her work as Chancellor?
3) How can we systematically change individual perpectives on data from “useless numbers” to “valueable insights”?
4) Can too much data be a problem?
References