Skip to content

Week 3: Biodiversity Data Cleaning

Emerging Frontiers Symposium: AI, Genomics & Community Science in Biogeography We will join first for this symposium from 10-11:15 or so, and then reconvene at 11:30 at the usual class zoom link.

Icebreaker

What is something you enjoy doing when you are not doing science?

Introductory business

  • Isaac will be giving the LSU Museum of Natural Science Seminar next week (09/25), so the status of class is up in the air at the moment. I am going to try to get the morning off to join for class but I'm still working on it. Also, if they have a zoom option I will forward this if you want to watch my talk.

    • LSUMNS Seminar (4:30pm Eastern) "Integrating eco-evolutionary models and machine learning for predictive biodiversity science."

  • The usual intro update exercise: Go to your JupyterHub and pull the latest version of the class github repository.

Commands to pull the latest version of the class repository
  1. Change directory to your local copy of the course repo
  2. Pull the latest copy of 'upstream' which is my copy of the class website
  3. Push the changes to your own github repo 
    cd ~/BIO597-SpatialBiodiversity/`
git upstream
git push
  • Review Assignment-02-pygbif.ipynb

Paper discussion

Core Questions

  • What can go wrong when using biodiversity occurrence records?
  • How should taxonomic, coordinate, temporal, and metadata uncertainty be documented?
  • How can bad spatial data produce convincing but wrong analyses?

Concepts

  • Occurrence records
  • Handling taxonomic uncertainty
  • Assessing coordinate uncertainty
  • Removing duplicate observations
  • Dealing with sampling bias
  • Standardizing metadata

Python Tools

  • pandas
  • GeoPandas

Applied Lab

Students download occurrence data for a species or taxonomic group and build a cleaning pipeline that handles missing coordinates, duplicate records, impossible coordinates, obvious spatial outliers, coordinate uncertainty, and duplicate specimens or observations.

Assignment

Remember: Next week Isaac will be in Baton Rouge. We will almost certainly still have class but it will probably only be the first 2 hours. More to come...

Complete and submit this assignment, including a data cleaning report with raw and filtered maps, a table of filtering decisions, and a short explanation of remaining uncertainty.

  • docs/assignments/Assignment-03-DataCleaning.ipynb

Paper discussion leader next week: elle! Please select a paper for the group to discuss before Sunday 09/20 6pm and send it around to the class email list: bio597-fall2026-group@maine.edu

Be prepared to lead a brief discussion on this paper next Friday 09/18, following the guidelines in the Paper Discussion section on the assignments page.