« BACK TO MAIN
Skip to main content Link Search Menu Expand Document (external link)
  • The premise here is: You have some data, you’d like to make it available over the Internet, how hard is this?
    • The good news is: You can be up and running with a toy example in a matter of a day
    • The reality is: Building a full-up data service you are happy with is likely to be something of a project
    • Therefore: Be sure to have a demand that justifies the time investment to build your supply
    • Let’s quickly outline three approaches to consider. We elaborate on the second and third approaches below.
      • First approach: Place the data in a logical structure on the cloud and enable your “customers” to read it
        • Simplest; but your data users need to build compute environments on the cloud
      • Second approach: Place the data on the cloud and build a data service using a Web Framework
      • Third approach: Place the data on the cloud and build a data service using Serverless functions
  • Web Framework approach to data as a service
    • A Web Framework is an assembly of code and libraries. We install it to publish a generic interface; and then we customize
    • A good Web Framework to read about is called Flask: It is not too complicated, nor is it too super-simple
    • A really powerful Web Framework, in contrast is Django; much bigger time investment needed to climb the learning curve
  • Serverless approach to data as a service
    • Serverless compute is a simple way to build some intelligence into a data service
    • Think of serverless as “code that magically runs and I don’t have to worry about the computer or the operating system”
    • To do: Provide the link to the re-tested serverless tutorial “Zero2API”
  • A blog on data science practice containing some useful links…
    • Data archival: Dataverse, Zenodo, FigShare, Dryad
    • Also links for: Research data, source code, workflow, executable publication, tools, licensing, privacy