Urban_Heat_Islands_demo

This repository is the Urban Heat Island Project demo, made for the purpose of PyCon Submissions as well as Geopython Submissions for the year 2022.

Setup the environment

Use the environment.yml file to install required packages.

conda env create -f environment.yml

Data Samples

https://archive.org/details/@sumedhghatage?tab=uploads&and[]=mediatype%3A%22software%22 Make sure you arrange the data in the given folder structure

Steps involved in the project

There are total 6 notebooks involved which takes into consideration Data preprocessing, Indices calculation, Spatial Merging, zonal statistics and Spatial regression model

1. Calculate satellite variables pre-processing Landsat 8 data.
2. Convert to time-based grids
3. Add non-satellite variables to grids
4. Merge satellite and non-satellite variables
5. Calculate derived variables: Shannon's entropy and building block coverage
6. Create spatial regression model

Quick Information Guide

Satellite Imagery Data - Landsat 8

With the presence of changing atmospheric conditions, there are images that get hampered due to the presence of the cloud cover. We must filter out those images on the basis of the scene cloud cover.

Scene cloud cover is nothing but the presence of cloud cover in the area of interest. The below image shows one of such scenarios.

The images selected for the modeling purpose through visual inspection and the metadata information for the image using USGS Earth Explorer

Zonal Statistics and Aggregation

Zonal statistics calculates or performs the statistics with the values of the value raster. Then get the statistics values for each zone based on the zone raster. A zone is defined as all the cells that have the same value in the input (the discontinuous cells of same value are accepted to be a zone).

Zonal Statistic as Table shows the result in Table instead of raster. Zone raster defines the shape, value and position of the zone, and the input value raster provides the original data for calculation.

Here the number of zones provided based on the grid size. The grid size was created using supermercado by mapbox at zoom level 18 (which is approx. 100m Grid ) All the zonal statistics for landsat data are calculated using the method above. The Morphological factors were then aggregated on the basis of each grid using mean, minimum and maximum values.

The zonal statistics for each year is calculated for each year from 2013 to 2020 for the summer months (From May to September every year.)

Spatial Regression

Spatial regression methods allow us to account for dependence between observations, which often arises when observations are collected from points or regions located in space.

Regression (and prediction more generally) provides us a perfect case to examine how spatial structure can help us understand and analyze our data. Usually, spatial structure helps models in one of two ways. The first (and most clear) way space can have an impact on our data is when the process generating the data is itself explicitly spatial