Guide to Building Data-Driven Organizations in the Public Sector

Discovery

Team 3: Cody Lundell, Christina Worden, Dan Gabiou

Topic Overview of Chapter 8: Discovery

This set of readings includes Chapter 2 of Social Physics: How Social Networks Can Make us Smarter, Chapter 8 of Reality Mining: Using Big Data to Engineer a Better World, and several discovery-related blog posts from the I Quant NY website. In Chapter 2: Exploration of Social Physics…, Pentland (2015) states that “the most consistently creative and insightful people are explorers” (p. 26). By explorers, the author means that good ideas come from collaborating to obtain different perspectives, enhancing the idea, and repeating until confidence is built. Or in other words, “exploration is the part of the idea flow that brings new ideas into a work group or community (Petland, 2015, p. 39). This concept could also be applied to gaining insight from big data and social networks, where “copying successful people, yields real rewards” (Pentland, 2015, p. 33). To ensure exploratory success (positive results from your collaboration efforts), Petland (2015) recommends: 1) diversity, 2) social learning (copying successful people/ideas), and 3) seeking out contrarian consensus (learning why groups of people are opposed to an idea) (p. 39-41).

In Chapter 8 of Reality Mining: Using Big Data to Engineer a Better World, Eagle and Greene (2014) show examples of how data visualization (such as timelines, maps, and graphs) may be used to help predict, prioritize, prepare for, and address social, environmental, political, and economic issues. Data visualization goes beyond gathering data and identifying trends, it’s compiling such data into comprehensive visual aids to help policy makers and the public understand the direness of the situation to help in decision-making.

The take-away from the I Quant NY blogs is the importance of having and sharing data with the public. The blog authors were able to use available data to review government practices and openly identify inefficient use of tax payer funds. By posting data online, the public has more access to validate government work, discover issues, and potentially enable corrective action or create added value.

In summary, our Discovery readings suggest that big data can serve many purposes and help improve our world in many ways. Step one is collecting, publicly sharing, and validating such data. Since we have data overload, it’s important to explore, collaborate, and challenge data to ensure accuracy of both the data itself and what it’s telling us. Lastly, data must be transformed in to easy-to-understand visualizations in order for such data to tell the full story and inspire action. Only then can we maximize the use of such data and more efficiently and effectively implement solutions.

Chapter Summaries

I Quant NY Blog Post Discoveries Summary

In the first blog post, Data Shows No Increase In NYC Plowing as Storm Picked Up, the author felt that inefficient snow plow operations were causing tremendous traffic problems, particularly during the evening rush hour. The author obtained raw snow plow/street centerline directly from New York City, depicted such data in GIS maps, and ultimately discovering that the city had not properly planned for a recent 5”-snow event, resulting in hour and a half delays on average, with some areas not being plowed at all.

In the second blog, entitled Open Data Reveals $791 Million Error in Newly Adopted NYC Budget, the author feels that the top line item in the NYPD’s FY2017 budget table ($791M for Protection of Foreign Missions) was in error. After doing a little exploration by reading the articles the author provides to support this claim and comparing NYCPD’s budget to Phoenix PD’s budget, I’m not sold that this claim is valid. To compare, Phoenix PD has had a $600M-$688M total operating budget for the past three fiscal years (https://www.phoenix.gov/budgetsite/budget-books/Summary-Budget-2018-19.PDF). The author compared the $791M figure to a $27M figure from a 2012 police financial statement, but that figure appears to be for a different program. Although it is questionable that this is actually a $791M error, transparency of data did help the author identify the need for NYCPD to give a more transparent and consistent naming convention for their Staff Salaries budget line item.

The final blog I read was Payer or Prayer – A Look at NYC’s $650M Property Tax Breaks Related to Religion. The blog author noted that in 2015, property taxes made up for 27% of the city’s $70B budget, yet it was very challenging to obtain property tax data from the city. After public citizens created an index from the city’s online property tax PDFs, it was realized that property tax exemptions totaled $12.9B, with Houses of Worship contributing $476M, which ipso facto is being paid by non exempt property tax payers (Payer or Prayer… 2016, https://iquantny.tumblr.com/).

Whether discovery claims are valid or not, the take-away from these blogs is the importance of having and sharing data with the public. At the end of the day, most government agencies attempt to be transparent by complying with Open Meeting laws and posting meeting minutes, but many committees are not well-attended and most overseeing boards are not technically savvy, relying on government data to be accurate. By posting data online, the public has more access to validate government work, discover issues, and potentially enable corrective action.

Reality Mining Chapter 8 Summary

Big data from demographics, mobile usage, internet/social media advertising and searches, tweets, and financial institutions are covered in Chapter 8 of Reality Mining: Using big data to engineer a better world. Fascinating examples of data convergence are examined which can impact everything from health, personal finance, national sentiment and poverty. Eagle & Greene (2014) make the point that data visualization and storytelling through timelines, maps, and graphs, can be used by policymakers and nongovernmental organizations to see impactful results.

Big data varies in its availability and timeliness. For example, census and World Bank data provide important insights into populations but the data refreshes slowly whereas mobile phone data refreshes nearly in real time. Combining the historic data from the census with the real time data from mobile phones allows for better understanding of population growth and control. Converging this big data together, it would be possible to better understand, for example, the population flows of slums and the price of food staples. These models could help predict population influxes and allocate the right resources in the right location at the right time.

There are some interesting and seemingly beneficial examples of using big data to better understand the flu season to track the internet searches relating to the flu or flu symptoms. This search could then be mapped to figure out where the flu is affecting people most in a country. Some examples are discussed in theory like financial data mining which is mostly controlled by the large banks. If this data was anonymized, it could help to anticipate economic disasters and help to develop data-based guides to improve depressed economies.

Exporation

Many “ordinary” people believe that “extraordinary” people are almost inhumanly intuitive, creative, and resourceful. The exploration concept would suggest another method for how great ideas and behaviors actually come to be. Exploration of ideas requires a few key concepts such as testing compelling stories, diversity, winning down ideas, social interactions, idea flow, and echo chambers.

Compelling stories. Compelling stories are the genesis of all great ideas or expansions of existing ideas. Individuals that explore usually always have a collection of ideas that they are testing out with people. The challenge is to test a wide range of people, skill sets, and perspectives. That is the only way to understand how effective the story is against reality.

Diversity. Diversity is crucial to successful exploration. Average performers will only discuss and bounce ideas with like individuals. This limits that applicability and susceptibility of their idea to change behavior. An explorative individuals continues to test their ideas.

Winnowing down ideas. As stated previously, individuals that explore always have a collection of ideas or stories. The process of testing against diversity is called winnowing. Just as an elegant blade’s edge can be honed to a hair-popping sharpness, sufficient winnowing will product a solid idea that is ready to integrate.

Social interactions. Integration of ideas within a social group has been around since the beginning of time. The most primitive cultures of any species still find success when social interactions are built. Great ideas are ones that flow well.

Idea flow. In a nutshell, great ideas will impact a significant amount of people’s behavior. Susceptibility to the new idea’s ability to change behavior is key in exploration.

Echo chambers. There are some dangers to be aware of with social interactions and people relying on others’ ideas. These are echo chambers or what are commonly known as fads and bubbles. Great ideas that are copied many times become viral and can create specific influxes of similar behavior. These can be dangerous because all fads and bubbles will eventually pop. Great ideas must stand the test of time.

Key Take-aways:

References