Today satellites provide a surface signature of the temperature with a high spatial frequency: ie a good horizontal resolution but a low vertical resolution. Thanks to the ARGO database collected by buoys making vertical surveys, one has precise but sparse knowledge of the vertical thermal structure of the ocean. The objective of this project is to develop a methodology to statistically combine additional information to obtain a 3D time series sufficiently resolved horizontally and vertically to follow the eddies. It is based on a regression classification method that allows both to classify the temperature profiles and to propose a linear model between the satellite observations and the information of the buoys according to its label.

🌏 A project for visualizing and analyzing global earthquakes (M ≥ 2.5, last 30 days) using a single R script that automates data download, cleaning, and plotting. Generates 15 sequential plots covering spatial, temporal, and statistical patterns, including regression analysis of magnitude vs. depth.

This project entailed building a machine learning classification model for predicting the presence of EV charging stations in a location based on socio-economic, transportation and geo-spatial characteristics acquired from public and government data. The model was then used to further explored areas of potential for new EV charge point installations using the unsupervised KMeans algorithms to cluster areas with favorable characteristics from the classification model. Raw data was preprocessed in order to ultimately train and test a Logistic Regression classifier and statistical testing of feature coefficients were done to confirm the factors that increased or decreased the likelihood of EV charge point installation.

This project aims to estimate building's height on historical maps based on the building's characteristics in the present time. Random forest regression are used to estimate the building's height, and the result will be visualized into 3D City Model Visualization that could be used for further spatial analysis.

GIS project - Use of geographically weighted regression and k-means clustering in R to investigate patterns of wellbeing in London in 2011. Includes use of shapefiles, spatial data frames and output layers for the Office for National Statistics

The final project of my graduate applied spatial statistics course in which I plotted choropleth maps, conducted Geary and Moran tests and constructed lagged Poisson regression model to examine the spatially-autocorrelated patterns of NYC policing stop rates

Self-contained project to demonstrate three spatial analysis concepts: Spatial Autocorrelation (Moran's I), Hot Spot Analysis (Getis-Ord Gi*), Geographically Weighted Regression (GWR).

Final project Assignment in Eric Huntley's Spatial Statistics Workshop. My project executes a spatial regression analysis on positive covid cases in US states for 12 months after the lockdown in March 2020. I control for independent variables such as population density, political leaning, time etc to predict covid cases.

Project using basic workflow and functionalities of spatial-regression-related R packages

Project for using deep learning for nonlinear regression with few variables. Spatial interpolation etc.

House price prediction using spatial clustering and geographically localized regression models. Made for the CAP5610 Machine Learning term project.

A mesh simplification strategy for non-planar spatial regression with a penalty term. This is a joined project for NAPDE and APSC courses.

This repository investigates spatial interpolation of sparse temperature sensor networks using regression–kriging. Using Heidelberg (Germany) as a case study, the project integrates environmental covariates (elevation, canopy height, and building height) into a RK framework to improve the spatial prediction of hourly mean temperature.

This project analyzes the impact of urbanization on land surface temperature (LST) in Mysore using remote sensing techniques. Key components include NDVI, NDBI, and regression analysis for spatial heat pattern assessment.

This project contains an analysis of Philippine Eagle habitat preferences using spatial statistics, Geographically Weighted Regression (GWR), and MaxEnt. The project explores the influence of environmental covariates, such as elevation, river proximity, and forest cover, on eagle presence.

A geospatial data analysis project exploring health risks like heart disease, malaria, dengue, TB, and child malnutrition using Python. Techniques include spatial clustering, regression, and vulnerability indexing to uncover actionable insights for public health planning.

A data-driven project that predicts Wisconsin county and tract-level populations by integrating geospatial, census, and land-use data using regression models. Combines machine learning and spatial analysis to uncover how area, housing, and land features influence population distribution.

This project focuses on analyzing a forest fire dataset through exploratory data analysis, feature engineering, and predictive modeling using both linear regression and random forest algorithms. The dataset includes details about forest fires such as spatial coordinates, month, day, weather conditions, and the area burned

2016-2023 Personal Portfolio: Urban strategy, regional economics, and China's industrial policies. Utilized ArcGIS spatial analysis, regression, LSTM forecasting in Ministry of Housing and Urban-Rural Development projects. Detailed insights into city planning, economic drivers, and national policies contribute to a comprehensive understanding.

Built a data-driven analysis of Chicago crime patterns using spatial visualization and machine learning. By combining GeoPandas-based geospatial mapping with KMeans clustering and logistic regression, the project explores where crime concentrates across Chicago and how community-level socioeconomic factors relate to overall crime severity.

This project addresses the challenge of predicting the precise $(x, y)$ coordinates of a single active pixel (value 255) within a $50 \times 50$ grayscale image1. While the input is image-based, the task is framed as a Supervised Regression problem to map spatial data to continuous numerical outputs2.

This is the final group project of the course "Environmental Data Analysis." We gathered the data from EPA and used R to analyze 19-year county-level particulate matter data over the U.S. to discover the temporal and spatial air pollution pattern. Both parametric and non-parametric trend tests were conducted. As for spatial analysis, kNN, Kriging, and local regression methods were tested to see if there is any presence of a specific spatial pattern for the particulate matter in the United States. The results suggested that PM pollution is more a local than a regional/national issue. In the second part, we linked the air pollution trend to the USDA government investment in energy efficiency to examine the effects on air quality improvement. The results showed that the investment may have partial impact on air quality, while it requires further investigation to distinguish the numerous potential influencing factors.

As the population increases, demand for food increases too, which has led to large-scale land conversion to improve livestock production in Colombia. Fulfilling these criteria of increasing demand in a sustainable way is a challenge and remote sensing data provides an accurate method to support this task. In this study, Planet Scope multispectral satellite datasets and coincident field measurements acquired over test fields in the study area (Patía) of September 2018 was be used. Fresh and dry weight biomass was calculated and forage quality analyses, crude protein (CP), in vitro dry matter digestibility (IVDMD), Ash and standing biomass dry weight (DM) was carried out in the forage nutritional quality laboratory of International Centre for Tropical Agriculture (CIAT). Field data was related to the remote sensing data using the random forest regression algorithm. R was required for the statistical analysis, to figure out the model performance for IVDMD, CP, Ash and DM. This project also investigated the spatial distribution of livestock which is affected by quality and area of potential forage zones. The R2 values of the regression models were 0.74 for IVDMD, 0.69 for CP, 0.38 for Ash and 0.49 for DM using a predictor combination of vegetation indices, simple ratios and bands.

Obtaining information about the conditions of the right-of-ways on brazilian highways represents a primary need in preventing irregular occupations on them. In order to minimize the risks associated with irregular occupation on the right-of-ways, resulting from anthropic processes related to land use and coverage, it is necessary to monitor the areas close to the highways. In this sense, mathematical models based on cellular automata can help to understand the impact of changes in land use and / or coverage and to predict future changes in the margins of highways. The study area comprises the BR-104 highway, located in the forest area of the state of Pernambuco, and aims to assess the urban growth of areas that are crossed by highways, generating an annual land cover scenario for the 10-year period. The modeling carried out in this research used land cover maps from the years 2000 and 2010, obtained through the Mapbiomas Project as input. Eight variables were used to explain the changes in land cover on the margins of Highway BR-408. For the calibration of the Model, the Artificial Neural Network and Logistic Regression method was used. The model was validated by the Kappa concordance test. The results show that the generated scenario presents an increase of 0.50% for the class “Irregular Occupation over the Right-of-Way”, in the period of 10 years. This increase has the contribution of cells that were in the initial stage as class Forest / Non-Forest Formation, Agriculture and non-vegetated area. The results of the modeling point to the growth of occupations in the peripheral areas of the cities cut by the highway under study, showing a spatial segregation.

Spatial Statistics Project about "Penerapan Model Regresi Spasial Dalam Menentukan Faktor – Faktor yang Memengaruhi Persentase Penduduk Miskin Menurut Kabupaten/Kota di Provinsi Jawa Barat Tahun 2021"

Spatial Statistics Project about "Analysis of Variables that Explain the Human Development Index in West Java in 2021 Using the Geographically Weighted Regression (GWR) Method"

Spatial Lag Regression for Housing Data (based on 4th year project)

Project exploring spatial mapping and regression of employment insurance benefits in Canada

Projects in spatial analysis covering demographic mapping, bookstore accessibility, urban green space equity, and spatial regression for environmental justice.

THIS PROJECTS COTAINS CODE STATISTICAL MODELS APPLICATIONS SUCH AS PCA, LINEAR REGRESSION, LOGISTIC REGRESSION AND MODELS FOR SPATIAL DATA