Reproducible research

This section discusses and illustrates practical aspects of data use, with an emphasis on the reproducible use of existing data.

In the first chapter we discuss general aspects of reproducibility. In the next two chapters we show case studies in which we show our attempts to reproduce published research using the data from experiments and surveys. The case studies illustrate useful data wrangling tricks. They also show that reproducing research with open data can be very challenging, and in some cases perhaps impossible. Sometimes the data is poorly organized or difficult to understand. In other cases the data published does not match the data reported. We hope that these examples will illustrate useful techniques, but also help in creating better organized datasets in the future.

The fourth chapter discusses a case study on aggregating secondary data into a new data product. If existing data can be repurposed this can increase efficiency by saving time and money that would be required for further or repetitive data collection — if that were even feasible to do.

The fifth chapter discussed general lessons learned and ways forward.