Description:

Geospatial representation are so prevalent in day to day life, such as even in simple travel related conversation to maps, aerial/satellite images etc. In digital era, geospatial data is extensively produced and consumed in ever growing proportion. Python with its free and open source libraries are giving wide variety yet simple and effective set of tools to visualise and analyse geospatial data. The current workshop is directed for beginners of Python programming language, who have basic understanding on computing and data formats. The primary objective of the workshop is to introduce and give hands on training on selected list of FOSS libraries for geospatial analysis. The workshop as a do it yourself fashion tries to solve two real world problems in Geographical Information System (GIS) and its geospatial data sources.

The workshop comprised of three components:

Component 1: Introduction

Python environment and work flow setup, an assisted task of setting up the Docker and Jupyter notebook setup. Setting up the Geographical Information System (GIS) environment with extended discussion. Setting up of GIS tools such as FOSS QGIS and Google earth. This component is comprised of four exercises. 1. Introduction to vector data, 2. Introduction to raster data, 3. binary and text file formats of geospatial data, 4. Introduction to tools of GIS, 5. Introduction to literal programming- Jupyter notebook

Component 2: work with vector data

Find characteristics of road network(type of road network, length of the type) within a 1X1 km grid. The data source is Open Street Map (OSM) road network data on a city level (60X60km size). This operation is operationally simple such as measure a line feature but computationally intensive as the operation comprised of geometry within operation on dense road network seen in urban setup. Libraries such as Shapely, Fiona, Geopandas and rtree index will be used for the fast processing of this operation. This component comprised of three exercises 1. Find distance between two points 2. Find distance between two points constrained by another vector 3. Find distance between large number of points in for loop

1. Geospatial data operations by Geopandas, Geometrical operations using Shapely
2. Map plotting of vectors using matplotlib toolkit Basemap, video from maps
3. Web mapping on vector data using Folium

Component 3: work with raster data

1. Basic workflow on satellite imageries using Rasterio, NDVI and other indexing
2. Map plotting of rasters using matplotlib toolkit Basemap
3. Web mapping on raster data using Folium
4. Accessing Google earth engine using its Python API

Note : There won't be any environment setup session separately, participants are requested to ensure below prerequisites are working properly.

Workshop Plan

2. Component 1- 40 minutes
3. Break- 5 minutes
4. Component 2- 50 minutes
5. Break- 5 minutes
6. Component 3- 50 minutes

Prerequisites:

1. The hands-on exercises are given as Jupyter notebooks in here
2. The docker imagery for the workshop is in dockerhub, version7 is latest, please update into latest image in the above link just before workshop
3. As the workshop depended on multiple libraries there are two options to carry out hands-on

Option 1

1. Use personal laptop and setup the docker with above image, use Jupyter notebook served from the docker, for a setup please go through this
2. Have latest clone of workshop docker image(version8) sudo docker pull airpollutionstudyindia/foss-pt-gsa:version8 and jupyter notebook repo of the workshop git clone https://github.com/nishadhka/FOSS-Python-GeospatialAnalysis.git --depth

Option 2

1. There will be Jupyterhub running on the workshop docker image available during workshop which can be accessed using participants' github account
2. Participants are requested to ensure fast internet connectivity
3. Please fill this form, requesting your github username to whitelist access to the jupyterhub
4. Have git clone of workshop notebooks from here git clone https://github.com/nishadhka/FOSS-Python-GeospatialAnalysis.git

Content URLs:

For workshop home here and here such as to get sample data, Jupyter notebooks, slides etc

For workshop slide is available here

Speaker Info:

I am a research associate at UrbanEmissions.info, PhD in Environmental Sciences from Sálim Ali Centre for Ornithology and Natural History, Bharathiar University, Coimbatore. My thesis was on "Particulate air pollution data for Coimbatore, India: real time monitoring and modeling with data-interoperability measures", which emphasised need of open and standardised real time data for better management of air pollution in the country. I got exposure into programming and open source software related to my research studies. I used free and open source libraries of Python for most of my studies, especially on geospatial data compilation, analysis and visualization. I have conducted workshop similar to the current one in Kovaipy, Scipy-India 2017.