Research Projects

Ploomber can help you manage your research project to enhance reproducibility and to run more experiments faster. Click here to see a sample project.

graph LR load[Load data] --> process[Process] --> exp1[Experiment 1] process --> exp2[Experiment 2] process --> exp3[Experiment 3] process --> exp4[Experiment 4] exp1 --> summarize[Summarize] exp2 --> summarize exp3 --> summarize exp4 --> summarize

Faster iterations

Thanks to incremental builds, you can make small changes to your data analysis code and quickly bring your results up-to-date, since Ploomber will only execute the code that has changed since your last run.

Run (and organize) more experiments

Ploomber allows you to run many experiments in parallel. You can parametrize pipelines to run the same code with different configurations.

Furthermore, you can quickly generate all the parameter combinations from a grid. If one machine isn’t enough, export to systems like Kubernetes or SLURM easily.

Example: Running a grid of experiments in parallel
pip install ploomber
ploomber examples -n cookbook/grid -o grid

Ensure reproducibility

Since Ploomber generates an output notebook (that may include any number of tables or charts) whenever you execute your pipeline, you can easily bookkeep the results of each experiment. Whenever you make changes, such executed notebooks from previous runs can help you verify the reproducibility of your results.

Share your analysis

Ploomber can orchestrate all your data analysis for you if you need someone else to run your code, all they have to do is execute the following command:

ploomber build