How can we improve Microsoft Azure Machine Learning?

Support JSON for input/output

Several custom algorithms I use are designed expecting input in JSON format. This is ideal for integrating with a web service, but unfortunately ML Azure doesn't seem to support it.

95 votes
Vote
Sign in
Check!
(thinking…)
Reset
or sign in with
  • facebook
  • google
    Password icon
    Signed in as (Sign out)
    You have left! (?) (thinking…)
    Lawrence Thatcher shared this idea  ·   ·  Flag idea as inappropriate…  ·  Admin →

    JSON parsing is not yet implemented as a separate module. As a workaround for JSON parsing, you could use Execute R Script in following manner:

    1. Compress the JSON file as zip package and upload to ML Studio as .zip format file.

    2. Connect the .zip file to Execute R Script rightmost port

    3. Execute following R script (replace “sample_json.txt” with your file name).

    library(RJSONIO)
    library(plyr)
    data.set = ldply(fromJSON(content=“src/sample_json.txt”))
    maml.mapOutputPort(“data.set”);

    Note that this assumes your JSON file holds an array of similar records.

    3 comments

    Sign in
    Check!
    (thinking…)
    Reset
    or sign in with
    • facebook
    • google
      Password icon
      Signed in as (Sign out)
      Submitting...

      Feedback and Knowledge Base