Urban mobility system optimization relies on accurately predicting human demand within space and time. Latest forecasting methods ranging from spatial panel models to Graph Convolutional Neural Networks (GCNNs) achieve high accuracy by modeling spatiotemporal interactions. Yet with technical advancements in the era of smart cities, bridging urban infrastructural complexities with complex technology often suffers from simplified sociodemographic considerations. Perhaps much of it is not "unexplained noise", but rather an oversight of multi-ethnic, -cultural, and -racial livelihoods. I challenge one of likely many sociodemographic simplifications, which is the reflexive assumption that time is a neutral, universal, and linear coordinate, a "clock time" that functions identically across all demographic groups.

Here I introduce sociotemporal polychronicity into the common necessity of transportation demand modeling, specifically bikeshare demand. Using Philadelphia's Indego system as a case study, I hypothesize that neighborhood racial composition influences temporal elasticity's strength and direction. By interacting demographic variables with temporal lags, I implement Michael Hanchard's concept of racial time, the "inequalities of temporality that result from power relations between racially dominant and subordinate groups", into the mathematical architecture of demand prediction.