The geo-challenge: objective

• Demonstrate scripting skills in a spatial scripting challenge
• Choose or come up with an original and novel topic
• Write code using packages you haven’t used before, make complex reusable functions with efficient code to handle spatial data
• Make code reproducible and document well, use project structure, make good use of git
• Create an interactive presentation and visualisation (e.g. dashboard, storymap, app, etc.)

Example geo-challenges

• See your own Exercise 1, and Brightspace for predefined challenge topics and datasets
• First identify a dataset that you can use (e.g. data available for download, etc.).
  • a small (< 1 GB!) dataset that you can analyse.
• A simple challenge so that you can apply and demonstrate geoscripting skills.
• Python, R, Bash, GEE, all, other programming/scripting languages are fine.

The geo-challenge: Deliverables

1. Project proposal: your specific challenge idea
2. A project, repository, that is reproducible and produces a result (map, analysis) for your presentation

• If you have external data, put the data files on MS Teams (NOT in your Git repository!)

3. An interactive pitch demonstration of your work where you explain:

• objective
• workflow
• results by using a clear visualization (demo, map, output, analysis)
• what have you learned?
• what would you do if you had more time (e.g. 6 months)?
  • Check the rubrics on CodeGrade!

Teams?!

• You will work in a team of four (2+2), but for the time being submit a topic in your current team
  • You can swap team members, but get consent from everyone involved
• If you already know what other team you want to work with:
  • Write a joint proposal and have only one group submit the proposal for the whole team of four
  • Remove submissions from CodeGrade if necessary!
• If you don’t know whom you want to be in a group of four with:
  • Write your own proposal, indicate whether it is on one of the predefined challenge topics or a custom topic
  • You will be merged with another group that proposed a similar topic
  • You will work according to the better proposal from the two

Proposal submission deadlines

• Submit your proposal before Wednesday September 30, 18:00
• Review the other teams’ proposals before Friday October 2, 17:15, via Codegrade
• If you get a NO GO, make changes before Friday October 2, 17:15 - all proposals must be approved then
• By then you will know which group you are working together with and will get your Project starter repository the following Monday. You can then start working on acquiring the data.

The geo-challenge: planning

• 8 days for project work (Mon-Wed) in a group of four
• We have Aurora B9010 for the whole two weeks, but check on https://te.wur.nl/ when!
• Last Thursday: make interactive demonstrations and give peer feedback
• Last Friday: interactive demo session

Final presentation day

• We will have two sessions for the interactive presentations using the side screens:
  • 10 groups will present (5 per room), 10 will watch, and then we swap
  • 5 minutes per demo pitch
  • You will be asked to present and answer some questions by the staff and commissioners
• After the presentations, we will go to a plenary session for awards, picture and feedback about the course

Questions?

Exercise for today

Modify the proposal starter’s README.md using Markdown syntax into a concise scientific proposal. One paragraph for each task is fine, can include lists. Include:

• Name for the 4-member team and list of members (if you already know this)
• Project title
• Topic number, or “own” if it’s your own
• Task 1: Research questions (Why?)
  • Define the aim or question(s) of your project, phrased as questions to answer
  • Be very concrete and specific, also quantitative
  • e.g. “How well can Random Forest classify trees, grassland and crops in the Gewata region?”

Exercise for today

• Task 2: Data (What?)
  • Provide a link to the data and describe the metadata (author, date, extent, resolution, etc.)
  • Can it be accessed via a script, and how (i.e. API or download link)?
  • Estimate how large the data set will be for your project
    • e.g. do this by a little test download of one file unless you decide to work in Google Earth Engine
    • Important here is to keep the data set small! e.g. below 1 GB
• Task 3: Methods (How?)
  • Which functionality or approach will you use to achieve this? Can you already name new packages you will want to try out?
  • What will be the result, output, and/or visualization?

Questions?