A composite indicator of liveability based on sociodemographic and Uber quality service dimensions: A data-driven approach

Abstract

This article aims is to create a liveability index that combines traditional population-based datasets and non-traditional data sources from Uber services, potentially supporting when pursuing new perceptions of a city. For that, a comprehensive study was performed focusing on Natal, a Brazilian urban center that has important socioeconomic and public/private transport services inequalities. In order to assess intra-municipal disparities, the units of analysis were neighborhoods and intra-neighborhoods (the Human Development Unit - HDU), which represents small homogeneous areas in socioeconomic terms. A data-driven workflow was conducted to process more than 4 million trips and to retrieve the Estimated Time of Arrival (ETA), considered in this work as a proxy for the Uber quality service. In order to evaluate variations, associations and spatial autocorrelation between ETA and sociodemographic features, different methods were applied, notably the Exploratory Data Analysis (EDA), the Exploratory Spatial Data Analysis (ESDA), and the Regression Analysis. From the results obtained by these methods and the application of a Principal Component Analysis (PCA), a Composite Indicator of Liveability (CIL) could be created and validated as a measure fitted to sociodemographic characteristics, as it describes high values in wealthier regions and lower values in the poorest areas of a city.