This repository contains a suite of R scripts designed for spatial data analysis, visualization, and modeling, focusing on socioeconomic and environmental datasets. Below is a description of each file and its purpose:
Handles data import and preprocessing:
- Loads census and house price data.
- Merges multiple demographic datasets.
- Creates
SpatialPointsDataFrameandsfobjects for further spatial analysis.
Performs basic exploratory data analysis:
- Summary statistics (mean, median, correlation).
- Violin plots and bubble plots.
- Linear regression and confidence intervals.
- Plot correlation matrix heatmap.
Conducts Kernel Density Estimation (KDE):
- Computes KDE for house point data.
- Visualizes density using
tmapandraster. - Extracts 75%, 50%, and 25% home range contours using
adehabitatHR.
Spatial integration and basic geographic operations:
- Spatial join between house points and census polygons.
- Computes average house price by area.
- Creates buffer zones around points.
- Visualizes spatial layers with various tmap layers.
Geographically Weighted Regression (GWR):
- Fits linear and GWR models.
- Maps local regression coefficients and R-squared.
- Includes side-by-side map grid visualizations for multiple variables.
Performs spatial interpolation:
- Generates Thiessen (Voronoi) polygons.
- Uses IDW (Inverse Distance Weighting) to interpolate house prices.
- Visualizes interpolation outputs as rasters and bubble overlays.
Choropleth mapping of census data:
- Maps variables like qualification using different color schemes.
- Adds legends, compass, layout adjustments.
- Demonstrates quantile-based classification.
Advanced mapping with point symbols:
- Dot and bubble maps of house price.
- Combined visualizations with census polygons.
- Shows proportional symbol maps with clean layouts.
Spatiotemporal air pollution analysis (Taiwan PM10):
- Processes hourly pollution data and monitoring station metadata.
- Constructs a
STFDFobject for space-time analysis. - Fits spatiotemporal variogram and performs kriging.
- Visualizes predictions and base maps.
Spatial autocorrelation and clustering:
- Global and local Moran's I computations.
- LISA cluster maps (High-High, Low-Low, etc.).
- Getis-Ord Gi* statistic for hotspot detection.
- Visualizes autocorrelation using
tmapandspdep.