Eikon for Developers: Can I pull historical data into my HTML5 App?

tristan.hale
tristan.hale
in TR Internal
1) Is this possible?
2) If not, is it going to be possible?
3) I am assuming this is possible if building a .NET app, right?
Thanks
Tagged:

Answers

  • Panit Wechasil
    Panit Wechasil
    We have a plan to start the work around the end of this month and should be ready a bit later. You can track the progress at
    http://iajira.amers.ime.reuters.com/browse/JET-478
  • Denis Dollfus
    Denis Dollfus
    Hi Tristan,

    In few words:

    1 - Possible, no
    2 - Planned, yes, JET will be extended to time series
    3 - Available in .NET Apps, yes

    Just curious to understand why an HTML5 chart doesn't meet your requirements: could you clarify your use case?

    Thanks
  • Denis Dollfus
    Denis Dollfus
    Panit was first :) Still, understanding Tristan's use case can help exposing the proper APIs.
  • bogdan.grosu
    bogdan.grosu
    Hi Tristan,

    Many thanks for the question. As Panit and Denis mentioned the story is as follows:

    - the .NET SDK currently provides access to the time series data through new, very easy to use API built recently by Denis;
    - the Web SDK currently does not provide access to time series data yet, but we are working on this and we are hoping to release a first version of this in August

    It would be great to understand your use case in more detail so if you could provide any information, that would be very helpful.

    Thank you,
    Bogdan
  • tristan.hale
    tristan.hale
    Panit, Denis, Bogdan - thanks for the answers. The question came up because I have built a simple portfolio app, and wish to add some VAR analysis. There are a few approaches one could take to this, but in general we need to calculate the means, standard deviations, and covariances of log-normal historical returns. Does it make sense to replicate the RHistory function? If, like me, you tend to prototype the logic in Excel before coding then this would be useful. T
  • Denis Dollfus
    Denis Dollfus
    Tristan, thanks for clarifying. I believe that we shouldn't do any math calculation in the browser because of a couple of drawbacks: - To my knowledge JavaScript V8 math calculations don't specify any calculation accuracy (at least ECMA specifications doesn't, eg Math.Pow is specified as 'implementation dependent'). - Unless you use HTML5 Web Workers, a long calculation (ie > 500ms) will make the application unresponsive by consuming CPU in the main thread. - Although very fast, JavaScript V8 is still about 2x slower than Java or C#, though this ratio can change a lot depending on the running code (see [benchmarks][1]). Calculating VAR server-side would be better, maybe on TRACS infrastructure. Contacting Gabriel Le Blanc or Philippe Lampietti may help. [1]:
    http://benchmarksgame.alioth.debian.org/u64/benchmark.php?test=all&lang=v8&lang2=gpp&data=u64
  • tristan.hale
    tristan.hale
    Hi Denis. Good stuff. I'll look forward to the history functionality, but take on board that JavaScript math should be kept to a minimum. Of course, whilst VAR is desirable, it is quite specific. What might be more useful is to have some JET functions which are a little more generic - for example MAX, MIN, AVERAGE, STDEV, CORREL - as applied to historical data per specified RIC(s), start date, end date, data item, periodicity, etc. The remaining math for a single position historical or analytical VAR would then be minimal. Portfolio (or incremental) analytical VAR would require a significant matrix operation on the covariance matrix. The Monte Carlo approach may be a step too far for the Web SDK!? T

Categories