Nipun Batra

I'd agree even more now! I'll give some thought and sketch out ideas here.

Would be great if someone from the community would take up this issue. This will help one get a good handle on nilmtk and also on different data sets (and...

@mridulmalpotra and @RishiBaijal : I'd like you both to get on this issue as soon as you've finished the previous assigned task (writing unit tests). We are planning to come...

Not really sure about the implementation, so can't comment at the moment!

I think that the preprocessing (shuffling included) could also be an argument in the configuration I had mentioned in: #753 I think Jack always intended building pipelines in NILMTK and...

>It was impossible for all chunks to be in memory. So what I implemented was loading parts of the raw mains and appliance data, preprocessing them accordingly and finally feeding...

@MingjunZhong @OdysseasKr In light of your above comments, how about a preprocessing stage that takes as input the dataset, buildings in each and start and end times in each, and...

I think if the .h5 is in the same format as the original datasets, then, it should be trivial for users to visualise the data. This is because NILMTK already...

Also, please see: https://github.com/nilmtk/nilmtk/tree/master/nilmtk/preprocessing and https://github.com/nilmtk/nilmtk/tree/master/nilmtk/stats

Thanks for the feedback guys! I'll make a start on this tomorrow as I need it for my research. I'll start with a hacky solution based just on Pandas data...