CSV's with the datasets from ComparativeDeerAnalyzer

[MichaelLantz]

Fixes mindsdb/lightwood#673

Please describe what changes you made, in as much detail as possible

• Adding ComparativeDeerAnalyzerdistribution csv from data sources mentioned in original issue

benchmarks

[George3d6]

Could you consolidate them or would the dataset not make sense then? And if the answer is yes, could you write an info.py file (similar to other datasets) describing the target and some appropriate accuracy function.

[MichaelLantz]

Hi @George3d6,

If you want to get a sense of what's was involved to get from the .DTA,DSC to .DAT and eventually to a CSV, the zipped xlsx file has embedded Power Query transformations to split data from the output structure (referenced in doc) CDA generated text files (.DAT's) but required adding headers and doing general cleanup. Similar to the DEERNet library used by spinach this data was also broken up into multiple chunks (in this case distribution and fit). I extracted a single CSV using a fit example which is what I think would be used as a target benchmark.

That being said the xlsx can output the remaining 2 chunks and any other sample (distribution and fit) if necessary.

Feel free to dig in. Hopefully this output is useful as a benchmark and is generally what we're after.