Efficient retrieval of broker estimates

I'm trying to retrieve historical EPS broker estimates for a number of US equities for the last several years. Conceptually, I am trying to do something like:

ek.get_data(rics, ['TR.EPSEstValue.value', 'TR.EPSEstValue.origdate', 'TR.EPSEstValue.broker_id'], parameters={'SDate':'2014-01-01', 'EDate':'2019-06-30', 'Period':'FY-6';FY2'})

Now the above query appears to be prohibitively expensive to run in one shot, and times out. But conversely, if I requested each ric/date/fiscal period individually it would be a prohibitively high number of requests. What is the best way to batch this request? For example, should I request this as a time series (perhaps one per ric), or should I request all the RICs and periods for a particular date and iterate over dates? And would it be more efficient to request multiple fiscal periods or to request those one at a time?

Best Answer

  • davidk
    davidk Contributor
    Answer ✓

    For future users, having now done the benchmarking I was hoping to avoid:

    1) Requesting additional fiscal periods inside an IBES request does not take appreciably more time (requesting 1 fiscal year and 12 fiscal years is the same amount of time when benchmarked).

    2) Adding additional dates to your request increases time by what appears to be a sub-linear factor. However, too many dates at once can cause a request to time out.

    3) Adding more securities to a request appears to be roughly linear time increase.

    So ultimately I settled on each request requesting 3 months of forecasts for 1 RIC for all available fiscal periods.

Answers