homecloud computingcloud data Processingcloud of thingsclound of networks
cloud data processing

Spatial Data Mining with PostGIS and Python

Duration: 5 Days

Course Background

The purpose of this course is to explore the use of Python data analysis and machine learning packages, as well as the use of Python interfaces with R packages in the exploration and mining of data held in a PostGIS database.

Course Prerequisites and Target Audience

Attendees are assumed to have a good working knowledge of PostGIS and Python as well as a basic knowledge of R.

Course Outline

  • KDD - Knowledge Discovery in Databases
  • Spatial data mining - the Process
    • Visual interpretation and analysis
    • Attribute query and selection
    • Generalisation and classification
    • Detection of spatial and non-spatial association rules
    • Clustering analysis
    • Spatial regression
  • Integration of GIS and Spatial Data Mining
    • Visualisation through GIS
    • Discovering association rules and minimum confidence thresholds
    • Mining for multi-level associations
    • Detection of spatial and non-spatial association rules
  • Data mining via scripts - available pacages and tools
    • Python scripting for R
    • Python scripting for PostgreSQL
    • Migrating geometry from PostGIS to GRASS GIS
    • Python scripting for GRASS
    • Python scripting and QGIS
    • Basic python numeric and scientific modules - Numpy, Scipy
    • Pandas - Python Data Analysis Library
    • Panda extension to handle geographic data - GeoPandas
    • PyBrain - Python Machine Learning Library
    • Scikit-Learn - Python Machine Learning Library
    • Mixing it up - Python + R + PostGIS
  • Making maps with Python
    • Shapely + Fiona + Mapnik
    • pymaps - Google maps for Python
    • PostGIS and GRASS
    • GRASS and Python
    • Integrating PostGIS with GRASS and Python
  • Spatial data mining with Python - tools, strategies and techniques
    • Data exploration
    • Decision trees and random forest
    • Regression
    • Clustering
    • Outlier detection
    • Time-series analysis
    • Association rules
    • Combining social network analysis with spatial data analysis