Abstract
In recent decades, the collection and use of geospatial data has generated interest in various fields, including economics, biology, and health. In our paper, we present our experience collecting, cleaning, and transforming such data. We then used it to classify Costa Rica's districts and cantons. The results were compared with other indices developed by government agencies in the country.
Keywords: geospatial data, google cloud, human development index, principal components, classification
Introduction
Every day, within the analysis of data, the use of geospatial data becomes more relevant, these allow us to show a more complete image of the events [4] in our environment. They are not always easily accessible, either due to the cost or access of them or the computational management.
Using geospatial data, we share our experience in the characterization of the districts and cantons of Costa Rica. This with the aim of being able to contrast our segmentation with existing groupings at the cantonal level, such as the human development index [3] or the consumer confidence index prepared by the University of Costa Rica [2]. To do this, we collect information related to social variables (schools, hospitals), economic variables (bars, restaurants, malls, banks) and religious variables (churches), this information was provided by Google Maps [1], we used a Google Cloud API, which returns the values found for a variable indicated in that area based on a center and radius. In our process, we made a circle overlay of the entire country. The process generated millions of records, which were cleaned and transformed by us, to build a summary table of information that served as input for subsequent statistical analysis.
A principal component analysis and a classification using k-means were performed. The results found allow us to observe the formation of classes of districts that are comparable with what was observed or that can be validated with instruments such as the aforementioned indexes.
5 April, 9:00 - 9:20, Room 13.2
XXXII Meeting of CLAD, Porto, 3 – 5 April 2025
References
- Google Cloud. Google cloud platform, 2024. Accessed on 8 January 2024.
- Universidad de Costa Rica (UCR). Informe del Índice de confianza del consumidor (ICC), febrero 2023, 2023. Accessed on 8 January 2024.
- Programa de las Naciones Unidas para el Desarrollo (PNUD). Atlas de desarrollo humano cantonal, 2024. Accessed on 8 January 2024.
- IBM. Geospatial data, 2024. Accessed on 8 January 2024.