Data Carpentry: Data Skills Training to Enable More Effective Research

Our increasing capacity to collect data is changing science. This is particularly true in genomics where, with high-throughput sequencing, data production is no longer a bottleneck. There is great potential for discovery, but we are primarily failing to translate this sea of data into scientific advances, because researchers are not trained in the skills needed for effective management and analysis. The question then becomes, in addition to scaling data production and computation, how do we develop and deliver training to scale data literate researchers? Course curriculums are slow to change, need qualified instructors and are already full. Short courses are oversubscribed and reach a limited number of participants. To provide scalable and distributed training, Data Carpentry develops and teaches domain-specific hands-on workshops in data organization, management, and analysis. This is a grassroots training effort developed by practitioners for practitioners, who identify core skills and collaboratively develop lessons. All lessons are open source, and workshops are taught by volunteers trained by the Software Carpentry Foundation. A particular focus has been on lessons for working with genomics data, designed for people with little to no prior computational experience. The workshop teaches in two days how to organize genomic data and metadata, use cloud computing, use the command line to run bioinformatics pipelines and analyze and visualize data in R – the full data lifecycle. Workshops are in high demand, but this model allows for scaling of training and teaches the foundational skills to get biologists started managing and analyzing their data effectively.