This year, we are partnering with the BPS Experience Lab, the education segment of New Urban Mechanics. This work has focused on visualizing data on student time on buses – length of trip, frequency of use, and conveying this information to headmasters through a dashboard using Power BI. Through that work (and some recent Globe articles), we learned that 10% of the BPS budget is spent on Transportation. And as identified by the BPS Long Term Financial Planning Initiative and the 10 Big Ideas to Unlock Resources for Student Success, addressing these transportation costs can free up funds to invest in student success.
BPS is hosting a challenge to better leverage technology to improve routes and bell times with the ultimate goal of reducing transportation costs. In TCE, we sit at the intersection of government, industry and non-profits and this kind of challenge is a terrific use of the data science capacity of the private sector to enable the public sector to better serve constituents and students.
We are proud to welcome John Hanlon and Will Eger as guest bloggers to tell us more about this challenge.
— Aimee Sprung, Civic Engagement Manager at Microsoft New England
Last Saturday, over fifty technologists, academics, and transportation industry leaders braved an early Spring snowstorm (only in Boston…) to join us to kick off the first-ever Boston Public Schools (BPS) Transportation Challenge — a data science competition, open to the public, aimed at improving Boston Public Schools’ bus routes and equitably and efficiently balancing our school start times. We are excited that this innovative public-private hackathon will help us reach — as BPS Superintendent Tommy Chang put it — “a technical solution to a technical problem, using data science to transform our district” in a way that provides the best outcomes for students and families. We are hopeful that this solution will come from one of the groups in attendance, including students from Northeastern, MIT, BU and Harvard, or from industry powerhouses like FedEx and Uber, but it could just as likely come from someone reading this blog! So… (shameless plug) if you have a knack for solving these kinds of problems then please roll up your sleeves with us and hack away!
The event helped remind us of a number of things. First, as our panelist Andy Rotherham — co-founder of Bellwether consulting — pointed out, “solving school district transportation problems is incredibly hard.” But as John’s remarks highlighted, it’s incredibly important not just for BPS but for Boston as a whole. Reducing the 45,000 miles our buses drive every day wouldn’t just allow us to reinvest in schools, it would also dramatically reduce our carbon footprint. Rebalancing our school start times could potentially free up funds for investment in the classroom while establishing school schedules that work better for families.
The event also illuminated the evolution of our bus routing system, something that is still a work in progress. Mike Hughes, the Assistant Director of BPS Transportation, reminded us of this when he said during the event’s panel discussion: “When Boston Public Schools began creating bus routes in the 1970s, we unrolled large and detailed maps of the city and used push-pins to mark bus stops and connected them with multi-colored string to form unique routes.” Needless to say, our routing and fleet management has evolved dramatically since then. Today, our 650 buses drive 45,000 miles a day and serve 25,000 riders at 231 public, charter, and Parochial schools.
Technology has played an increasing role in planning these routes. Push-pins and strings have been replaced by routing software and digital maps. However, our software still can’t solve this puzzle without placing a significant burden on our excellent drivers, who often have to operate on inefficient routes, or on our talented transportation staff, who need to troubleshoot and fine-tune the computer-generated routes each summer.
And why is that such an issue? As research into the Traveling Salesman Problem has found, as the number of stops increases the permutations of possible routes grows on factorially (n! – that is, possible permutations increase faster than exponential growth). Therefore calculating the optimal solution by brute force becomes impractical after about 20 stops. And we have 5,000 unique stops, at which our buses stop about 20,000 times per day (the same stops often serve multiple buses).
Things get even more complicated when you factor in the many “rules of the road” that we have to consider when routing. These rules establishing ride-time maximums, bus-stop placement rules, and so on, quickly make this problem nearly impossible to solve.
But there’s hope! With the tremendous advances in digital mapping, the rebirth of the Traveling Salesman problem in academic circles, and the sheer growth in computing power, we believe that now is the time to try to solve this historically unsolvable problem. We think that there just might be someone out there who can develop an algorithm that creates a more optimal solution to both routes and school start times.
As we think about our wish list, we know that this algorithm must be adaptable. We want to better understand the true costs of our various policy choices regarding walk to stop distances, ride times, and student assignment. Given the interconnectedness of our system, we’ve learned that seemingly small changes can snowball into large cost changes. What we want in the end is a tool that not only reliably automates efficient bus routes but also acts as a calculator of sorts, quickly and agilely determining the system-wide impact or cost of various policy scenarios.
Lastly, this is a technical challenge – but one with a very real human component. For 25,000 students, their school day begins when they step on the bus. Therefore this challenge isn’t just about improving efficiency. It is also about ensuring that our students reach schools safely and on time. It is about ensuring that schools start and end at times that work for more families. It is about reinvesting in our schools.
We hope to see your entry in the our routing challenge – make sure you don’t miss the 4/30 deadline and visit our website to learn more!
John Hanlon has served as the Chief of Operations for Boston Public Schools since July of 2015. Prior to becoming COO, John worked for the City of Boston as the Commissioner of Property and Construction Management where he oversaw the management, maintenance, and operations of City Hall and other municipal facilities across Boston. He previously served as Chief Operating Officer for Scholar Athletes, a nonprofit that supports public high school athletes and was the longtime Executive Director at the Dorchester Educational Enrichment Program, a nonprofit that offers mentoring services for middle-school youths. John is a proud Boston Latin School graduate and Dorchester resident, where he lives with his wife and four children. He holds an MBA from Duke University and a BA in journalism from Boston University.
Will Eger is a Strategic Project Manager in Finance for Boston Public Schools, where he works on developing and implementing the district’s Long Term Financial Plan. Prior to this he was in Parthenon’s education practice and was a high school math teacher in Philadelphia. He has written on education for The Atlantic, Ed Week, the Huffington Post, and Higher Education in Review as well as a full length book on the Tea Party. He has an A.B. from Harvard College and a M.S.Ed from the University of Pennsylvania.
Tags: Boston, Boston Public Schools, Boston Public Schools Transportation Challenge, BPS, BPS Experience Lab, BPS Long Term Financial Planning Initiative, cambridge, Hackathon, John Hanlon, microsoft, Microsoft New England, New England, New Urban Mechanics, Power BI, Will Eger