Document output data format

[lazd]

Document output data format

[DavidGeeraerts]

Yes that's me [jumping up and down]!

[DavidGeeraerts]

I'm looking into whether to make this an ISO 19115:2013 metadata record or use metadata-json.
This is after all, Vector Spatial Data --> ISO 19115:2013

[lazd]

@hyperknot, thoughts on this?

[DavidGeeraerts]

I'm heavily swayed by data.gov Open Data Standards:
Project Open Data Metadata Schema v1.1

I'm somewhat spinning my wheels trying to find a good metadata-json editor. I'd rather not have to edit in raw. I'm looking...

[DavidGeeraerts]

This is exactly what I was looking for and need!
mdEditor
Let's just say it's in the works after being stalled for the day. :-(

[DavidGeeraerts]

@lazd I've validated the workflow with mdEditor:

1. Import the mdeditor metadata file
2. merge
3. edit
4. export
   Think this workflow will work as a collaborative metadata project.

My question: should I upload the metadata.json file to the coronadatascraper repository?
I would think so, just want to check with you, since that's not exactly where the data is.
Should another repository be created just for the data?
(organization: CoronaVirusDataTeam)

[lazd]

@DavidGeeraerts how often will this metadata file change, what does it look like, and how should we provide it?

[DavidGeeraerts]

Getting an official metadata record for data.gov is going to take (me) a while. Went back to creating the metadata.md file so you can post on [your] website. I've started a Github organization called "CoronaVirusDataTeam" --that was before the whole Coronavirus Response Data Team)-- and have a repo for docs. In there is the draft for metadata.md. This is a lot of work and I can only do it in chunks as time permits.

Can you provide me with (or update the document):

featureId:
description:

aggregate:
description:

I think I can do the rest of the CSV entries in the format I've laid out. Not sure about the JSON files, haven't looked closely at them.

[lazd]

• featureId - The ID of the feature this location references (index in features.json's FeatureCOllection)
• aggregate - Whether this data source provides aggregate information (i.e. this county-level data was scraped from a state-level source)

[hyperknot]

I recommend keeping this simple for now, that's the key for adaptation and future contributors.

I think the ideal timeseries would look like this:

{
  "id2:ITX": {
    "dates": {
      "2020-03-17": {
        "cases": 2,
        "deaths": 0,
        "recovered": 0
      },
      "2020-03-18": {
        "cases": 2,
        "deaths": 0,
        "recovered": 0
      }
    },
    "source_url": "https://github.com/CSSEGISandData/COVID-19",
    "population": 68414135
  }
}

Where each date string is a ISO 8601 date and each id is a Country Level ID.

Addressing counties, etc with an id is not a simple task, but I'll think about it. Maybe we can do something like id4:"US-CA-ABCDE" or something. I don't think there is a consistent system for county short codes.