Data management is the most cumbersome and time consuming task for an Operational Research expert. Connecting an optimization model to data and modifying the data can take up to 80% of a data scientist’s worktime! This is a big issue, especially now that access to data-science libraries is so easy.
To address this issue and free up the time of such specialists, we created a small light API – called doopl – to embed OPL models in the Python ecosystem.
With very few lines of code, you now have the ability to easily use pythonic data structures to read and write tabular data with OPL.
It also simplifies a lot optimization workflows that require multiple solves with data changes.
Here is a non-exhaustive list of data management possibilities:
- CSV/Excel files and SQLite databases with Pandas.
- SQLAlchemy connection to handle databases.
- Tuple lists coming from a forecast done with scikit-learn or any ML library.
You can find examples showing all the capabilities of the library on github: doopl-examples