I am also trying to figure this out! I need to know if the release is going to be before or after hours.

Hi Matias, I was able to figure this out. You can retrieve the data using: EarningsCalendar.previous_calendar_time.latest. Then you can a python .loc to get a specific stock in your pipeline output's earnings time.

One thing that I noticed is that the data was just recently changed such that if a stock reported yesterday after market, then previous_calendar_time = 0. Before, previous_calendar_time = 1 for a stock reporting yesterday after market. Are these type of data changes announced somewhere? I wasted many hours trying to figure out why the heck my algorithm's backtest from last week varied from the backtest yesterday despite having the same code. Additionally, when deployed live, it would be very dangerous to be have earnings date data change on you since earnings are such volatile events. I really think Quantopian should send an email or notification if a dataset you purchased is modified.

Hi Sofyan,

We did recently apply a change to the way that calendar times were applied to the BusinessDaysUntilNextEarnings and BusinessDaysSincePreviousEarnings factors. Previously, the factors were treating all announcements scheduled on day N as the same business day, regardless of when they occur. We didn't think that this made sense because on day N, you wouldn't be able to distinguish between events that had already occurred (in the morning) and events that are scheduled to happen after the market closes. As such, we changed the design such that anything occurring after market open is considered to be scheduled on the next trading day - essentially treating the business day for day N as 4pm on day N-1 to 4pm on day N (Eastern Time).

The change hasn't yet taken effect in research, but it should be this week. Initially, this is why we delayed the announcement, but I see your point regarding live trading with the dataset. I just sent out an email to subscribers.

I apologize for the lack of communication.

Jamie, thanks for the reply. This update is definitely an improvement to the data and I appreciate the accountability.

Thank you very much Sofyan! I hope you can answer one more question I have...

The strategy I'm working on takes two different approaches to trading when earnings are released before market open vs after market close.

earnings_type = EarningsCalendar.previous_calendar_time.latest  
    if earnings_type == "Before Market Open":  
        earnings_type = 0  
    elif earnings_type == "After Market Close":  
        earnings_type = 1  

In the notebook examples the values for the variable calendar_time is either one of those two strings, but I'm not getting any matches with this. How would I go about classifying it now?

Hi Matias,

EarningsCalendar.previous_calendar_time.latest is a Classifier. I would recommend going through the Pipeline Tutorial to understand how these work.

In general, this is a good example for assigning labels to stocks based on the previous calendar time.

from quantopian.pipeline import Pipeline  
from quantopian.research import run_pipeline  
from quantopian.pipeline.data.eventvestor import EarningsCalendar  
from quantopian.pipeline.factors.eventvestor import (  
    BusinessDaysUntilNextEarnings,  
    BusinessDaysSincePreviousEarnings  
)
from quantopian.pipeline.filters.morningstar import Q1500US

def make_pipeline():  
    amc = EarningsCalendar.previous_calendar_time.latest.eq('After Market Close')  
    bmo = EarningsCalendar.previous_calendar_time.latest.eq('Before Market Open')

    pipe = Pipeline(  
        columns={  
            'after_market_close': amc,  
            'before_market_open': bmc,  
        },  
        screen= (Q1500US() & (amc | bmc))  
    )  
    return pipe  

Then, you can attach this pipeline to your algorithm and look at the pipeline output to get stocks by previous_calendar_time.

Thank you very much! This has been very helpful.