Spatial statistical analysis

• Main Author: Oreena Raveendran
• Other Authors: Fedor Duzhin
• Format: Final Year Project
• Language: English
• Published: Nanyang Technological University 2024
• Subjects: Mathematical Sciences; Spatial analysis
• Online Access: https://hdl.handle.net/10356/175637
• Institution: Nanyang Technological University

The Covid-19 pandemic situation was dire in New York City (NYC), prompting immediate measures to mitigate the transmission. These Stay-At-Home (SAH) measures altered the geographical distribution and rate of crimes. In this study, the Inhomogeneous Cross L-Function and global and local spatial autocorrelation methods – Moran’s I test, Geary’s C test, Local Moran, Local Geary and Getis-Ord – were used to analyse this change. From the Inhomogeneous Cross L-Function, we found that there is spatial randomness between 0 − 50m, clustering between 50m − 170m and inhibition beyond 170m for the two spatial processes: Covid-19 and Crime (7 types). Furthermore, using global spatial autocorrelation methods, we deduced that Covid-19 did not affect the overall spatial distribution of crimes. Lastly, the local spatial autocorrelation methods allowed us to understand the change in spatial patterns across NYC for the 7 types of crimes from the pre-pandemic period, to during pandemic and also the post-pandemic period. We also analysed the variation in formulas and results from these local spatial autocorrelation methods. Overall, the results drawn from this paper suggest that Covid-19 did not have as significant of an impact on crimes despite NYC being the hardest hit city in the US from this virus.