Dataset and the results in Tab.1 in the paper

[StarBlue98]

Hi, thank you for sharing this valuable repository and your work on deep learning for CO₂ emissions estimation.

I noticed in dl-co2-pp/Simulated/org_model_train_eval.py that the comment says: "This file is on the replication of the results from the Lippendorf or Boxberg dataset." I have a question regarding Table 1 in the paper/repository:

• Could you clarify whether the result for the Simulated Data CNN method in Table 1 corresponds to one specific dataset (Lippendorf or Boxberg), or is it an average over both datasets? This would help me better understand how to interpret the reported performance.

Additionally, I understand from the description that the Satellite Data was collected from a proprietary dataset owned by IBM.

• Is there any plan to open-source this dataset in the future?

• If not, could you kindly suggest any publicly available alternative datasets that are suitable for training and evaluating deep learning models for CO₂ emission estimation from satellite or simulated imagery?

As a graduate student who is new to the field of deep learning for CO₂ monitoring, I am still learning about the available datasets and preprocessing pipelines. Any guidance or clarification you could provide would be greatly appreciated.

Thank you again for your contribution and for making this work accessible!