Data Days 2018 at MassMutual

Ryan Hutchins
Saturday 02 March 2019

During the period from July 24-27, 2018, which occurred during my second month at MassMutual, I had the opportunity to participate in an organization-wide effort called Data Days for Good, hereafter referred to as “DDfG.” DDfG is a four-day event wherein MassMutual employees drop their regular work and engage full-time in well-scoped, small-scale technical projects to enable charity and community organizations to better serve the community. This year, there were a host of interesting projects aimed at contributing to worthy causes.

The projects undertaken by MassMutual’s teams during DDfG 2018 were:

image alt <>

DDfG was without question the most rewarding workplace volunteer opportunity in which I’ve ever participated, and I think it speaks to both MassMutual’s organizational commitment to its vision to “bring financial wellbeing to all Americans,” as well as to the ingenuity and philanthropy of its employees that the event has been such a success.

The chief contributor to the program’s success is its organizer, Christine Pfeil, with whom I recently had the opportunity to speak about the program’s historical development. DDfG began simply as Data Days back in 2014. It was a biannual event, then, in which individuals from across the organization formed small, cross-functional teams to work on projects of personal interest. Often these projects involved creating a tool to fix some operational problem for which they may have had a temporary workaround and, despite any annoyance caused by the workaround, wouldn’t have the time to redress in daily work.

The teams at MassMutual benefited from the fruits of these mini projects, and the participants really enjoyed the opportunity to step back from daily work, meet some new engineers, work on new problems, and build tools to smoothen the daily workflow of their comrades. Somewhere in all of this fun, the idea emerged to use the collective competencies of MassMutual directly for the good of the community. So, MassMutual reached out to local non-profits and charity organizations to find out how we could leverage our technical expertise to help them with their humanitarian missions, and Data Days became Data Days for Good, with the first event in the summer of 2017, and it has continued to grow now through its second year.

The project that I worked on personally was to create a web service that enables participants in CirclesUSA’s program to visualize their position vis-à-vis the Financial Cliff Effect. If you are not familiar with this effect, the Financial Cliff Effect is the name given to the situation wherein families at the low end of the income spectrum can experience precipitous drops in the value of services available to them as a consequence of very small increases in income. The result of this drop can mean that a family that was able to meet all of its monthly expenses while its earner(s) were making a particular, lower wage could suddenly fall very short of its monthly expenses because they received an increase in wages. This would seem to incentivize exactly the opposite behavior that fiscal policy should seek to incentivize. CirclesUSA’s goals for the tool are twofold: (1) to empower participants in their program to be aware of their financial standing with respect to the financial cliff, and (2) to create a series of case studies of their clients’ data to bring to policymakers so that they might enact less draconic rules for public welfare programs.

The reason that the Financial Cliff Effect occurs is that, as a result of poverty, the government provides services to families such as housing assistance, food assistance, and child care assistance. The amount of each service that is provided is a function of a family’s income, as well as of the number of children in the household. These services are necessary living expenses, and they each carry a certain monetary cost to replace if purchased without assistance. For a single income household, a jump of $1.00 in hourly wage equates to an increase of about $160.00 per month in gross income. If this wage increase pushes the household over the financial cliff, they could, for example, lose child care assistance valued at over $1000.00 per month, which is a very large net loss and could drive them well below the net zero income level.

The vision for our project was provided by Vince Gonzales from CirclesUSA, who has a tool that provides some functionality for creating graphs tailored for individuals’ financial situations. The tool, a snapshot of which can be seen in Figure 1, is a website that provides a form for collecting a person’s personal financial data. It then uses this form to generate a graph, whose x-axis is hourly wage and whose y-axis is monthly buying power, which is defined as the total of their net income plus the monetary value of the services that they receive from the government. The service lays several lines out on the graph, including a horizontal line for monthly expenses, and a plot of the buying power as a function of hourly wage.

Vince wanted our team to create a website that had functional parity with the original website, but that was built on modern web technology with a clean, well-engineered interface, and whose underlying code was production grade. Production code integrated with a modern web technology stack would ensure that the project remains functional, extensible, and maintainable, even when responsibility for the code base is eventually handed over to CirclesUSA. This extensibility will be very important as CirclesUSA continues to expand the tool’s regional coverage. Each state has its own set of rules for its benefit programs, and Circles’ current tool only covers the state of Michigan.

The clean interface would assist CirclesUSA participants in entering their data correctly and understanding their results, thus enabling them to visualize both their overall financial situation and their position releative to the financial cliff without getting into the minutiae of their monthly spending habits. To actualize this simplification, we planned to separate the interfaces for data entry pertaining to different benefits programs into different HTML forms, and also to remove an existing form which contained fields for highly detailed data about monthly budgeting, which were originally grouped together in a single form. The goal behind these changes was to make the experience less overwhelming and time consuming for the user. To further the goal of making information available in a simple and timely manner, we also aimed to create user sessions through which we could save user data so that it doesn’t need to be input each time they log in, which is a drawback to the existing version.

The specific set of objectives that we set out to achieve with the project were:

  1. Create a new website that has functional parity with the existing website.
  2. Use an extensible architecture for the new website so that its features can be easily expanded in the future.
  3. Provide the ability to store and retrieve user sessions.
  4. Provide the ability to page back and change the data in the current session.
  5. Provide the ability to print the results on a single sheet of paper, as well as sharing and exporting the data in common formats.
  6. Clean up the user data gathering forms.
  7. Create a mobile device friendly web app out of the website.

The project was fun and also very ambitious for our team, which was comprised two data scientists, a business analyst, and two back-end programmers who had, at the project’s start, never used HTML, CSS, or Javascript, the technologies on which the web’s front end and, subsequently, the CirclesUSA website must be built. The team spent the first few days learning the ins and outs of front-end web technology, and at the suggestion of knowledgeable members of our team at MassMutual, the JQuery library for Javascript, which anyone who uses Javascript regularly will tell you is a time investment that will pay back exponentially. During the four days of DDfG, our team managed to learn HTML, CSS, and Javascript to the point where we could build a website that reached parity with the original DDfG website, and that implemented extensible, well-written code on modern web architecture. So, we completed goals 1, 2, and 6, and we set ourselves up to implement the remainder of the desired functionality, and it’s a testament to the team’s talent that they were able to build an interactive website from scratch and integrate several independently-developed pages into a cohesive whole while having zero web development experience in just four days.

After the event, the team unanimously voted to continue working on the CirclesUSA project independently and on their own time, and I would like to single out Emad Taliep, Deepshikha Adhikari, and Cheng-Yin Eng for their outstanding and continuing effort on the project. Cheng Yin is also doing an excellent job providing transparency and communication to CirclesUSA, as we try to balance developing the new website with our lives and daily work. As of now, we are scheduled for a two-day DDfG reboot in December, during which we hope to have completed all but step 7. Stay tuned for an update at that time.