Clarification on Data_Orientation

[regnans]

• Documentation reference (URL): https://github.com/ess-dive-community/essdive-file-level-metadata/blob/master/flmd_quick_guide.md#data-orientation
• Submitter: Kim
• Justification (why is this change necessary?): I am a bit confused

I suggest the following changes: I'm not 100% sure on choosing whether my data is "horizontal" or "vertical". If each row contains a unique data point (e.g. different sample), then I would describe that as vertical, as new data additions are added in a vertical sense. However, that contradicts with the description that suggests that data organized in rows are "horizontal". I think I am misunderstanding what is meant by "data organized in rows". Can you clarify in the description which term should be used to describe data where each row is a unique data point and different columns contain? Perhaps a way to clarify would be to align with the terminology in the csv reporting format. If a file has column names, it is "this orientation", and if it has row names then it is "that orientation". thanks.

[regnans]

Ah, I see this is defined in the csv format, as:
The orientation of the Column/Row Names in the Data Matrix could be presented:
1. Horizontally with Names at the top of each column or
2. Vertically with Names at the start of each row.

Therefore, my suggestion to clarify the Data_orientation section is to note that what is being asked for is the header orientation. So my data with unique records in rows, has horizontal headers.

[wavingtowaves]

Thanks for highlighting how the description for this part of the FLMD is confusing. We kind of bounced back and forth between what the clearest name for this particular element would be.

@regnans, @tvelliquette, and @charuleka please let me know if my proposed new definition for this element of the reporting format is clearer, and also respond with any edits.

Current definition: Describe how the data are organized within the data matrix. Choose between "horizontal" (i.e., data are organized in rows) or "vertical" (i.e., data are organized in columns).

Potential new definition: Describe how the data are organized within the data matrix. Choose between "horizontal" (i.e., data are organized in rows) or "vertical" (i.e., data are organized in columns). In the case of "horizontal" data, header names start each row. In the case of "vertical" data, header names are at the top of each column.

[regnans]

Thanks Rob. The additional sentence in the new definition will be enough to clarify, but I think it is currently the reverse of what is stated in the csv format (if the names are on each column, the data is horizontal - see my comment above).

[wavingtowaves]

@regnans @tvelliquette I made revisions to both the CSV / FLMD formats so that we are aligning on terminology. I agree with you Kim, I think we had the language in the CSV format reversed or at the very least it was a bit confusing how we were talking about "data orientation" and "row/column name orientation"

Here's the updated portion of the CSV format. What do you both think? https://github.com/ess-dive-community/essdive-csv-structure/blob/master/csv_quick_guide.md#column-or-row-name-orientation

[regnans]

Apologies, I dropped the ball on this one. I think we are getting close, but can clarify just a little more.

I like the updated description in the .../csv_quick-guide.md#column-or-row-name-orientation. However I don't think the "reporting format statement" for that element is accurate. something like, "Describe orientation of the data matrix" would be better? (or the same as in the FLMD, "Orientation of tabular data".

And updating https://github.com/ess-dive-community/essdive-file-level-metadata/blob/master/flmd_quick_guide.md#data-orientation with the "potential new definition" suggested above will make it clear.