-
Notifications
You must be signed in to change notification settings - Fork 1
Home
This content is aimed at members of Earth Lab. We use this wiki to document best practices and use of CU and Earth Lab resources and assist with onboarding new team members.
New to Earth Lab? Check out our onboarding page for help getting started! You may also find it useful to explore our best practices wiki as you begin projects in Earth Lab.
-
Our Portfolio of Active Tasks that links to our Trello board with transparent task tracking across the breadth of our active projects and efforts.
-
Data Tools and Software projects includes open source data tools (not quite software) and packages developed by our team and also those projects that receive contributions from the Analytics Hub (these contributions are useful for reporting purposes)
-
Drones includes what Earth Lab is doing in the drone space, what drones we have, how to access them and use them.
-
Trainings includes trainings that Earth Lab hosts.
To better track the compute infrastructure development and curation by our team, we have the following pages:
-
CU Cloud Storage that talks about the different data storage options available, best practices for getting the most of each, and what to do when you leave CU Earth Lab.
-
Scalable Compute with HPC and Cloud computing, including CU High Performance Computing, Google Earth Engine, and Amazon Web Services to enable scalable compute on BIG data that cannot be processed on your local desktop without access to these resources.
-
Deploying code from Jupyter Notebook on HPC Learn how to convert your Notebook code to an executable and run it on the university HPC. Also useful for learning general HPC skills.
-
Docker containerization technology to enhance workflow portability.
-
Using Very High Resolution Satellite Imagery includes how to access different resources from Digital Globe, Planet, NAIP, Worldview, etc.
The Analytics Hub engages Earth Lab members and affiliates in incubator science projects:
- Understanding and predicting wildfire extremes: https://github.com/mbjoseph/wildfire-extremes
This is a collaborative project with team fire and others to try to improve upon previous efforts to model wildfire extremes.
- Deep learning to identify sonar anomalies: https://github.com/mbjoseph/sonar-anomalies (private)
This is a collaboration with Carrie Bell and Kris Karnauskas to use deep learning to automate anomaly detection in NOAA sonar data.
- Supporting wildfire emergency response with social media: https://github.com/mbjoseph/emergency-tweet-filter (private)
This is a project with Jeremy Diaz and Lise St. Denis aimed at automating Twitter filtering to deliver relevant tweets to emergency managers in real time as wildfire disasters unfold.
Incorporating Planet Dove Imagery for Scientific Investigation: https://github.com/joemcglinchy/<link_coming>
This project seeks to understand how we can incorporate the multi-system and multitemporal image data acquired by Planet Dove satellites for scientific investigation along with other well characterized satellite image data sources such as Landsat.
Impervious Surface Mapping in Urban Areas using DigitalGlobe WorldView Satellite Imagery: https://github.com/joemcglinchy/<link_coming>
This project investigates mapping urban cover type using DigitalGlobe multispectral satellite data acquired through their Geospatial Big Data (GBDX) platform.
Tracking wildfire spread using IR satellites and machine learning:
This project is prototyping an approach to predict wildfire spread using GOES-16 ABI infrared observations and a machine leaning algorithm detecting hot spots and screening for clouds. The goal is to demonstrate a use case for the Tools-Applications-Processing (TAP) Lab as part o the CU-Boulder AIA Grant.