Replica, a transportation analytics company, provides data and insights to help their customers better understand the built environment. Replica continually looks for ways to make its data outputs more accurate. Their customers consist of a mix of policymakers and planners in the public sector, as well as private sector clients who want to incorporate geospatial insight into their strategic and city planning.
High-quality data is vital to planners as they make important local policy decisions, such as how much parking they need to provide, where to put the next subway stop, or where there is demand for more housing or commercial office space. However, access to this data is usually slow, limited, and expensive.
Replica runs a seasonal, high-fidelity simulation that accurately represents the population and its travel patterns for the entire country. They collect and ingest historical and near real-time data to create sophisticated models of current and future mobility patterns. Replica’s customers rely on this information to improve the planning and monitoring of transportation and land use systems in cities nationwide.
Gridwise Analytics’ gig mobility data was chosen by Replica and integrated into their model as an additional source of ground truth in January 2023, using data from fall 2022.
Recognizing the challenge
With a focus on economics, infrastructure, and transportation, the models Replica produces require data about traffic patterns and people’s choices about how they will get around.
The percentage of trips involving Transportation Network Companies (TNC) has propagated substantially over the last decade. These companies include app-based platforms, such as Uber, Lyft, DoorDash, and Instacart, which move people and the items they want through independent workers known as gig drivers. While TNC growth has been substantial, it’s been hard to find data and mobility insights that effectively shed light on the new patterns of movement.
However, as the percentage of trips involving TNCs grew, more data was required to inform Replica’s models.
Replica relies on data sets sourced from observation of traffic counts and the types of activity involved. It consists of auto and vehicle counts in aggregate and specific vehicle types. Also included are public transportation and TNC/taxi counts, and public datasets.
While observation and public data yielded basic TNC information, Replica’s team had a desire for ground truth —namely, TNC data activity and mobility insights that could effectively capture the scale and detail required to depict the expanding TNC activity across the US accurately. This would allow Replica to more effectively address its customers’ concerns about how TNC activity affects issues such as parking, increased congestion on the streets, and how taxes for TNC services might be structured.
Replica knew that it was searching for additional data that tracked TNC activity, but sources for such data were limited. Not every city reports TNC data, making it difficult to cover Replica’s vast customer network. Thus, it was necessary to do a lot of extrapolating, which is not beneficial to the model.
The solution? Truth in data
Screenshot showing network volume by household Income
When Gridwise Analytics mobility data caught Replica’s attention, they realized they had come upon a data source that would be valuable in informing Replica’s model with more detailed TNC activity.
The data Replica uses is from several input sources. In aggregate, it is used to train and calibrate the model. Replica uses Gridwise Analytics as ground truth data to perform both of these operations. These include delivery data, rideshare data, gig driver trip and location data, TNC data, gig economy data, as well as others.
Holding out a portion of the data makes it possible to determine if the models match what is happening in real life. Gridwise Analytics offers additional ways to discern patterns and make predictions. “We always knew we needed TNC data, but before Gridwise Analytics, there were only limited sources out there,” they say.
Accurate data results in impactful models
Using Gridwise Analytics data and mobility insights, Replica can refine its perception of TNCs’ role in transportation throughout their customer population. Replica develops trip and fare models that enable their customers to estimate costs per trip and show the numbers and percentages of trips that TNCs handle.
Screenshot showing hourly breakdown of trips by destination, filtered by TNC mode
Replica found Gridwise Analytics easy to incorporate into their simulations. The data is delivered in a ready-to-use and well-structured format.
Replica is confident that Gridwise Analytics has helped improve what they do. Integrating Gridwise data has allowed Replica’s already powerful model to be adjusted so it more closely aligns with what they expected to see given the rise of importance of the gig mobility sector.
What’s next for Replica?
Replica intends to continue building on the use cases for Gridwise Analytics data within their models. This includes potentially expanding into incorporating “points of interest” data sometime in the near future. Gridwise Analytics mobility data would provide further granular insight into traffic data, congestion data, and travel behavior around airports, merchant locations, events, and other important aspects of transportation analysis and planning.
The impact of delivery trips will contribute more detail into Replica’s powerful models. Delivery also affects factors that concern Replica’s customer base, such as parking and congestion. Granular data describing these trips will give Replica even more valuable information to offer regarding transportation planning.