Back to DrivenData | Blog

Use of ArcPy library

Are we allowed to use ArcPy library for geospatial data preprocessing, as its paid and closed source?

@rszostak Thanks for reaching out about this. Per the competition rules solutions need to be made available under an open source license to be eligible for a prize. It sounds like this source would not fit that requirement.

Hope that helps!

1 Like

Hey @glipstein, can you please clarify this point if the entire tool chain must be open source or just the interface on top? The reason that I ask is that I’m primarily a MATLAB user and while I’m very happy to open source my code on top, I clearly can’t the underlying engine.

And looking further through your link to the competition terms, under the logic that the entire tool stack must be open source AND that I then have to hand over license to my entire source code that is just impossible. Imagine the following scenario: I’m using Python and I want to use numpy. Under the logic of the entire stack I’d have to hand over the license to both numpy and to Python itself (as well as scipy, gdal, etc) in addition to my source code that sits on top.

On the other hand the specific section of the clauses are: “source code used to generate the Submission”. Which seems to fit with @rszostak and my question that we hand over our code that sits on top of an engine.

Hi @pbrady - Thanks for the question. Python, numpy etc. are open source so that’s not an issue. You are correct that MATLAB wouldn’t be allowed here.

Hey @glipstein, I am working with Peter on this challenge and I am very sad to hear of this restriction, and honestly also a bit disappointed. From the guidelines we assumed it would be enough to provide our code under the MIT license, regardless of which language it was written in.

It also raises some more questions, e.g. will solutions using Anaconda be eligible? Also, will you check all dependencies of each submission to ensure that all used packages and tools are really available under the MIT license?

I feel that as long as a solution provides good results and is scientifically sound and reproducable, it should not matter which language or framework is being used.

We’ll discuss if we want to continue to explore the challenge anyway and see how high we can place on the leaderboard even if our approach is not eligible for a price. It is a very interesting topic and I think the challenge is very well done, I really appreciate the focus on interpretability and explainability. :slight_smile:

Best, Christoph

According to rules, looks like the solution dependencies could be under the permissive open source license, not necessary under MIT:

A Submission will be ineligible to win a prize if it was developed using code containing or depending on software licensed under an open source license:

@ckammer Thanks for following up! We certainly hear your perspective. The standard rule is that solutions must be able to run with dependencies that are freely and openly available to all participants. Unfortunately a tool that requires a paid license doesn’t meet that criteria.

As you mentioned, you are certainly still welcome to participate even if your solution is not prize eligible.