unsw-cfrc/teaching-geocomputation

Geocomputation module used in Geography BA/BSc at King's College London

Practical responsibilities!

• Add 30 minutes to lectures for 'talking through the practical'

• Add questions into the practical to test higher-level knowledge/understanding

• 1: Code Camp Recap [JM]

  • Scripts vs Notebooks
  • Errors
  • Data Types & Input
  • Operators
  • Comments
  • Conditionals & Logic
  • Some basic loops

• 2: Functions & Packages [JR]

  • Lists
  • Dictionaries
  • More loops [Lists of Lists; Dictionaries of Lists]
  • Use a lib to read a remote file
  • Use string.split to parse CSV data
  • Use a lib to parse CSV data
  • Create a function to read a remote file
  • Create a function to parse CSV data
  • Calculating values derived from CSV data (using function on LoL)
  • Calculating values derived from CSV data (using function on DoL)

• 3: Working with Data [JM]

  • Methods (for Pandas)
  • Link back to Week 2 and DoL
  • Pandas syntax for columns (both '.' and foo['bar'] and why they exist)
  • Describe, summary, etc.
  • Introduce string.format()
  • Max/Min/Range and other stats functions (e.g. IQR)
  • Also an opportunity to introduce reading .gz/.zip files directly in pandas
  • Look at basic pandas plotting functionality (lead into Seaborn in Week 4)

• 4: Visualising Data [JR? Complete]

  • Using pandas with header data
  • Statistics as judgement, not truth -- plotting as first step on this path
  • Seaborn
  • 3D plots
  • [Interpreting plots and other summary metrics]
  • Saving a plot
  • Automating analysis (loops but over data series now)

• 5: Assessed Notebook [JR]

  • Assign reading for following week: Openshaw 1998 and Wyly 2014

• 6: Working with Subsets of Data [JM]

  • Non-spatial Joins - join air quality data to existing LSOA data
  • Using the index [from Week 4]
  • Renaming columns (show how this was done in Data Loader nb)
  • Finding values, grouping by values (for Boroughs) [from Week 4]
  • Sampling data [from Week 4] -- will just highlight quickly
  • Matching on parts of a word, extracting parts of a word [from Week 4]
  • Exercises: Initial exploratory analysis of pollution data

• 7: Transformation and Standardisation [JR]

  • Thinking in 'data space' [From Week 4]
  • What counts as extreme?
  • Finding outliers
  • Residuals (first exposure to this)
  • Simple Transformations??? [from Week 4]
  • Simple Standardisations??? [from Week 4]

• 8: Making a Map [JR]

  • Objects and methods [copy from Spatial Analysis Week 2? 3?]
    • (use geopandas to anchor this)
  • PySAL and loading shapefiles
  • Look at impact of transforms on understanding of map
  • Joins? (non-spatial already done in week 6)
    • Illustrate using Airbnb raw data (also gives geopandas whirl)!

• 9: Correlation and Residuals [JM]

  • Builds on standardisation and normalisation from Week 7
  • Possibly use scipy here since it has rank and Pearson correlation, and is a very well-used lib in the real world
  • Plot residuals on map and in graph!
  • Seaborn (can include r^2 and rank correlation)

• 10: Aggregation and group-by [JM]

  • Brings in issues of scale and, implicitly, MAUP