Leveraging AI/ML to Improve Public Transit Equity, Access, and Operational Efficiencies

This article was originally published in Urban AI

Public transit has struggled to retain passengers in North America for many years, while in Europe it has proven to be a fundamental public good. This divergence in the role of public shared mobility versus private individualized mobility has reached a tipping point during the COVID pandemic.

While many private shared mobility services began to quickly expand across North America and Europe, public transit was caught in a difficult position. The battle lines were drawn, and many public authorities viewed private mobility services (scooters, bikes, ridehail, etc.) as a direct threat to ridership and revenue. However, things began to change with the onset of the pandemic.

COVID Mobility Landscape

Over the past two years, we witnessed a complete transformation of the movement of people and goods throughout urbanized areas and regions. What was the typical nine-to-five/Monday through Friday suburban/CBD commute has been turned on its head. Our daily patterns and routines were severely affected by public health policies that encouraged social distancing, working from home, and staying local.

These policies resulted in multi-peak commutes and ridership surges throughout the day, versus the duality seen pre-COVID. This disruption affected public transit probably the most of all modes (public and private). As a result, what started in early 2020 as an 80% to 90% drop in ridership, has slowly and steadily returned to normal, with numbers now equal to late 2019.

However, this return to normal has not been without its complexities and externalities. Public transit has indeed been able to regain passengers, but there are many considerations that public transit agencies and operators need to take into account if their service is to be sustained and repositioned to be more "user-centric."

Passenger Demand and AI

COVID-19 has resulted in a decreased number of passengers. Many public transport systems have adjusted their timetables to accommodate fewer passengers, but we are now moving into a growing market. Asistobe offers prediction functionality estimating the next period’s ridership, the passenger distribution across the network, and recommends headways to handle the change. Asistobe lets you optimize costs both for the short- and long-term.

Asistobe's short term predictions/algorithms of passenger distribution would have guided any PTA through COVID and now through post-COVID rebuild. Our long-term predictions ensure the public transport system actually handles the real transport demand as efficiently as possible.

Data Driven Decisions

Asistobe’s goal is to enable small to medium cities, without the financial muscles to invest in data science, to get access to a platform with advanced AI/ML algorithms. And in turn, empowering them to make data-driven decisions.

We also believe that smart use of data, and then also from several data sources, can help make public transport more efficient, such as our partnership with Telia to demonstrate and showcase the use telecommunications data for better understanding of real public transport demand, optimization, predictions, and planning.

Designing, improving, and optimizing a public transport network is an insanely complex task. At least if you are truly committed to transport as many people as possible, using a limited pool of resources. Asistobe’s tool provides you with AI-based functionality assisting you in making the right decisions for your city.