There is a separate directory for each week's work. As we progress, suggested answers will be uploaded.
The foundations folder contains some warm up exercises to transition on to the course. There are two folders within:
- basic_features
- iterating
The basic_features folder contains a set of scripts introducing the basic Python features (data types, loops, plotting etc.). Other than 06_objects.py, which will be covered in week 2, these should already be familar to you and are included here for revision.
The iterating directory contains some exercises to practice using loops and work up to moving around a raster dataset.
Week 1 covers:
Aspects
- Github version control and code repository
- Computer basics
- Revision of loops and file I/O
Algorithm
- Introduction to algorithm design: Finding minima and sorting
Contains two scripts to demonstrate reading from a text file and writing to a text file. The data folder contains the data for the file reading example.
Contains a text file with some sample data to practice reading with the fileIO code.
Contains an example solution for a simple sort algorithm. Note that there are multiple sort solutions, such as bubble sort, and then many more complex and efficient algorithms.
Week 2 covers
Aspects
- Using the command line to make programmable programs
- Objects and classes
Algorithm
- Binary search: Loop and recursion
Contains an example of the main block in order to ease importing code in to other programs
Contains two example python files, which can be used to alter the behaviour of a program at run time. This allows you to create a single python program and then reuse it with different input files, options etc.
commandExample.py: Minimum workable example of a command line
commandLineIllus.py: Illustrates the common command
The options for commandLineIllus.py are:
-h, --help show this help message and exit
--input INNAME Input filename Default=test.txt
--lat LAT Latitude in degrees Default = 5
--max_vcf MAXVCF Maximum VCF value to use Default = 100
--useSnow Use snow switch Default = False
-p [POW_BEAM_LIST ...], --power_beams [POW_BEAM_LIST ...] Track numbers of power beams Default = 5 and 6
Includes a script with a simple example of an object; a grouping of data and functions.
Contains some text data files for use in this week's exercises.
Contains a piece of code demonstrating the suggested use of document string comments.
Contains a suggested answer for week 2's algorithm. A suggested method using both looping and recursion is given.
searchObject.py: begins an object for sorting data
searchObject_withplot.py: the above with an plotting function added
binarySearches.py: contains suggested answers for binary search by loop and recursion
finishedQuartiles.py: uses the above to find quartiles in a sorted dataset
makeData.py: makes data for testing algorithms
randomWages.py: generates random wage data for testing algorithms
Week 3 covers
Aspects
- Geospatial packages: pyproj and gdal
- A note on function input/output
- Function fitting
- A mention of pandas
Algorithm
- Douglas-Peucker line generalization
Contains examples of document strings.
Shows an example of using the dgal package to reproject raster or vector data.
Shows an example of fitting a function to data.
Demonstrates the difference between variables that point to an array and copying a whole array.
Shows an example of reading data into a pandas object.
A function to find the orthogonal distance between a line and a point, to be used in the Douglas-Peucker line generalization algorithm.
Contains an example solution for the Douglas-Peucker line generalisation.
Week 4 covers
Aspects
- Geospatial data formats; HDF5
This folder contains some starter code to handle HDF5 formatted data from the LVIS lidar.
lvisClass.py - The base class to read LVIS data into RAM
lvisExample.py - An example of using lvisClass.py
processLVIS.py - Inherits from lvisClass.py and adds methods to extract ground elevation
lvisCompleteExample.py - example solutions for the week 4 tasks
Takes the following input parameters
--input INNAME LVIS input filename
--outRoot OUTROOT Output filename root to use with graphs and DEMs produced.
Week 5 covers
Aspects
- More geospatial packages: Geopandas
- RAM management
- A brief introduction to machine learning
Algorithm
- Machine learning; getting data into the machine
This folder contains an example of using geopandas to read a csv file containing geographic data, sorting and reprojecting it. Note that it makes use of the shapely package to define the geometry type within the geopandas array. Here Point is used as this is vector data. This stacking of packages can be difficult to follow, but the answers can be found on the documentation and stackoverflow. The latter is sometimes more helpful for specific tasks, as the official documentation shows all possible uses.
week5/geopandas/geopandasExample.py
This script has the following command line options:
--inEpsg INEPSG Input EPSG code
--outEpsg OUTEPSG Output EPSG code
--input INNAME Input filename
--output OUTNAME Output filename
To illstrate RAM management, a stack of rasters will be processed. Some starter code is available in:
week5/rasters/readTiff.py
This uses GDAL to read in a raster file, has a space to perform some data manipulation, and writes out the result.
This section gives examples for linear and polynomial fitting, to illustrate the components that go into machine learning. The code includes
linearFit.py # linear fitting example
linearFit_answer.py # linear fitting wth task answers
polyFit.py # polynomial fitting example
polyFit_answer.py # polynomial fitting with task answers
aic_answer.py # polynomial fitting with AIC task answers
A simple example is available in
week5/random_forest/randomForest_example.py
It has the following options:
--input NAMEN Input data table
--trainFrac TRAINFRAC Fraction of data to use for training
--nTrees N_ESTIMATORS Number of trees
--max_depth MAX_DEPTH Maximum branch depth
A more complex example is available that brings together all parts of the course. Note that this is not working as yet due to a bug in the data reading. This code is taken from Euan Mitchell's dissertation repository
week5/random_forest/randomForestLaSelva.py
The code in these repositories make use of the following packages and is all in python3:
numpy
matplotlib.pylot
h5py
pyproj
gdal
rasterio