Dima German

Short answers to frequently asked questions.

What is the DataHub project

Full answer is here: https://datahub.io/docs/about

Shortly, DataHub is a place where people can:

  • find a lot of high-quality data
  • store their own data
  • share data with other people (colleagues, customers)

What do you mean, saying "data"?

In the narrow sense data is just files with records, e.g. CSV, xls, etc… In the broad sense data is a organized set of information. In that case we call it Data Package, DataSet, Package Also data is the name of our command line tool, that used to work with data packages and Data Hub.

So you can use the data program to get a data package with data files from the data hub.

What are 'Data Package' and 'Dataset'?

Data Package or DataPackage or Dataset or Package are is the same thing in our context. It is a way of "packaging" up data - put the set of data files together with a special "metadata" file (also called 'descriptor'), that describes:

  • name, title, source of the data set
  • list of "resources" - data files presented in the package
  • each resource file structure
  • license

Data package also has a README.md file, that contains a general description of the data package for humans.

This links would help you, too:

What is the descriptor?

The descriptor is the central file in a Data Package. It provides:

  • General metadata such as the package's title, license, publisher etc
  • A list of the data "resources" that make up the package including their location on disk or online and other relevant information (including, possibly, schema information about these data resources in a structured form)

A Data Package descriptor MUST be a valid JSON object. (JSON is defined in RFC 4627). When the descriptor is available as a file it MUST be named datapackage.json and it MUST be placed in the top-level directory (relative to any other resources provided as part of the data package).

Read the full descriptor specification here

How to store my data on the Data Hub?

Install our data command line tool.

# go into the folder with your data
cd path/to/my/data

# create a data package
data init

# login on the DataHub (once)
data login

# push the data package
data push

# follow the online link `data push` has printed

Also you can read this guides:

Questions about the showcase page

Q: What is the showcase page? A: It is a web-page that presents one particular data package, stored on the DataHub, e.g. http://datahub.io/core/finance-vix

Q: Will keywords property be used? A: keywords property from the descriptor are not used on the showcase page ATM

Q: Will description property in datapackage.json be used to populate the Showcase description? How does that relate to the Readme section?

  • Yes, the showcase page description is populated from the description property in the datapackage.json.
  • If the description property is not provided - it is populated from the Readme section.
  • If the Readme property is not provided - it is populated from the README.md file from the package folder, when you run data init

Q: Will homepage property be used? I'm using it to link back to the original dataset page on UKDS.

Q: Will additional admin tools be available for the Showcase page? A: we are going to implement 'delete' and some other functions, and probably we'll make an admin page for these.

Q: Currently page is 'unlisted', is there a way to change this to 'public'? A: You have to republish your data with data push --public command

Q: Are my datasets listed anywhere, I can't see my 'unlisted' datasets on my profile page or dashboard. A: We have fixed this bug, so: You can see your 'unlisted' datasets on the profile page and dashboard.

Can I get an organization account/username rather than my personal username?

Yes, please contact us via [email protected] or https://gitter.im/datahubio/chat

Is there a way I can upload a dataset and be part of an organization?

This feature is in the roadmap: