The RAMP software packages were developed by the Paris-Saclay Center for Data Science (CDS) to connect data science to domain sciences. The original aim was to enable efficient collaborative prototyping of machine learning workflows in order to solve the data analytics segment of high-impact science problems. We then realized that it is equally valuable for training novice data scientists, networking and obtaining data on the way data scientists work and collaborate with each other.

The RAMP software has been used in many data challenge events aiming to solve predictive problems in science domains ranging from medicine to astrophysics. Data challenge events generally result in a model with significant predictive improvement over the baseline, which is very useful to domain scientists.

RAMP usage

Education

• Data Camp Masters 2 course - part of the Data Science Masters at Institut Polytechnique de Paris (IPP) and Computer Science Masters at Université Paris-Saclay.
• Data Science and Artificial Intelligence Masters program at Ecole Centrale de Lille.
• Data Science Professional education programme at Ecole Polytechnique.
• Machine Learning for Finance and Economics course at Université Panthéon-Assas.

RAMP challenge events

• SPIPOLL project - with the French National Museum of Natural History (read more at this blog by one of the participating start-ups).
• El Nino challenge - ran at the fifth Climate Informatics workshop in Boulder, Colorado.
• Macroeconomy Surrogate model challenge - with Paris School of Economics.
• Cancer rate prediction - with Epidemium.

Notable RAMP challenges

• Variable star classification
• El Nino temperature prediction
• Autism Spectrum Disorder detection (Public event)
• Pollinating insect classification

A complete list of existing RAMP challenges can be found on the RAMP starting kits Github repository.

Contributors

Contact

The CDS is able to host a limited number of events, challenges, or course projects, organized outside the CDS. Contact us if you are interested.