Awesome

labeller

A platform for collecting data for training and validation machine learning models used to map land cover. labeller was designed to interact with an ML model within an active learning framework.

This version evolved from the original version called DIYlandcover (Estes et al, 2016), that was designed to connect to Amazon's Mechanical Turk workforce. It was re-engineered into mapper, a standalone version with its own worker interface and (simple) task management system. labeller is a lighter-weight version of mapper, renamed to better describe its purpose.

Overview

Documentation is in the process of being updated. Here are a few pointers to some of it.

• A description of labeller's components, as originally designed.

• A description of labeller's database.

• How labeller and learner interact in the active learning process.

Building

• Building a labeller instance from scratch

• Cloning an existing labeller instance and using that to create a new instance

Citation

Estes, L.D., Ye, S., Song, L., Luo, B., Eastman, J.R., Meng, Z., Zhang, Q., McRitchie, D., Debats, S.R., Muhando, J., Amukoa, A.H., Kaloo, B.W., Makuru, J., Mbatia, B.K., Muasa, I.M., Mucha, J., Mugami, A.M., Mugami, J.M., Muinde, F.W., Mwawaza, F.M., Ochieng, J., Oduol, C.J., Oduor, P., Wanjiku, T., Wanyoike, J.G., Avery, R. & Caylor, K. (2021) High resolution, annual maps of the characteristics of smallholder-dominated croplands at national scales. EarthArxiv https://doi.org/10.31223/X56C83

Acknowledgements

The primary support for this work was provided by Omidyar Network’s Property Rights Initiative, now PLACE, with initial support from NASA (80NSSC18K0158), the National Science Foundation (SES-1801251; SES-1832393), and Princeton University. Computing support was provided by the AWS Cloud Credits for Research program and the Amazon Sustainability Data Initiative. Azavea provided significant contributions in engineering the connections between labeller and learner.