Estimize n days before release date

I just saw u have to pay for the dataset, so im sorry for posting

Hi Nico Xenox 

Did you pay?  Did it work better?  I subscribed to the estimize data for a month, but I had no idea how to code it, so would really like to try ideas and I would re-subscribe if I could get stuff to work, lol.  Please keep sharing.

Hey Newoptionz,

I found a better way for my problem without having to pay for a dataset. I just simply add the symbols to the notebook that I need the future earnings dates and import them via objectstore into the algo. If you're interested I can provide a example ;) 

sry for not replying but I had to change everything for it to fit my research. 

Yes Please.

I just re-subscribed to the estimize dataset, still trying to figure out how to use it.  Seems to me I have to build a list to track the past earnings and the price affects before and after earnings and see how these correlate to the various estimize prediction differences.  All without viewing the real data, lol.  

Nico Xenox 

I would say another method would be to track the avg number of days between earnings for a company and then use that as the future estimate of the earnings release.  Most companies do not even say when they are going to release earnings.  

Traders also watch the overall volatility of the weekly options market for spikes in the volatility, these signal most probable earnings weeks.  Maybe there is away to test that in a Jupiter notebook. 

Hey Newoptionz,

For this method you will first have to select the tickers you want to trade and after that import them into jupyter notebook:

1. Select tickers(via Hand or universe) save them into the objectstore 
    

#saves a list of tickers into the objectstore
import pickle
ex = pickle.dumps(self.tickers)
self.ObjectStore.SaveBytes('12843618/ticker', ex)

2. Import the tickers from the objectstore into the notebook

ticker_list = pickle.loads(bytes(qb.ObjectStore.ReadBytes('12843618/ticker')))

3. If you save the same tickers multiple times into the list check the list and delete them

symbols=[]
for i in ticker_list:
    if i not in symbols:
        symbols.append(i)

4. Request dates(NOTE: it will take some time)

from IPython.display import clear_output
self.list = {}
i = 0
for x in symbols:
	
	print(len(symbols)-i)
	i = i+1
    try:

		print(x)
        sym = self.AddEquity(x).Symbol
        date = self.GetFundamental(sym, "EarningReports.FileDate", datetime(2015,1,1), datetime.now())
        date = date.drop_duplicates()
        for x in date[x]:
            
            value = x.to_pydatetime() 
            self.list.setdefault(sym.Value, []).append(value) 
     except KeyError as err:
         symbols.remove(x)
         
     except AttributeError:
         self.list.setdefault(sym.Value, []).append(x)
     clear_output(wait=True)     

5. Save the dictionary into ObjectStore and import it into your algo

#Save
import pickle 
serialized_dict = pickle.dumps(self.dict)
self.ObjectStore.SaveBytes('12741724/dict', serialized_dict)

#Open in algo
dict = pickle.loads(bytes(self.ObjectStore.ReadBytes('12741724/dict')))
self.future = dict

6. Adjust dates and Symbols (You can not save Quantconnect symbols into the objectstore → Extract symbol values and rename with quantconnect symbol)

for x in self.future.copy():
	a = self.AddEquity("%s" % (x), Resolution.Daily).Symbol
    self.future[a] = self.future.pop(x)
   	self.RemoveSecurity(a)

7. delete first values of each symbol because that date already happened

for item in self.future.values():
	item.remove(item[0])

8. Last step.. not sure if it is necessary but I import every year a new set of symbols and have to adjust the dates dictionary to delete the already happened values.

i = 0
for x in self.future:
	i = 0
    for a in self.future[x]:
    	i = i+1
        if self.Time < a:
           i = i-1
           break
    del self.future[x][:i]

Hope it helps ;)

Thanks, looks complicated, I'll give it a go.

Nico Xenox Well, I need help.  I open up a new test.ipynb and then everything works, but it seems to only get the dates for one symbol and I don't know how to see the dict_list with the dates that are filtered. 

qb = QuantBook()
from IPython.display import clear_output
import pickle
tickers = ['PBBF','OXY','MOS']


ex = pickle.dumps(tickers)
qb.ObjectStore.SaveBytes('12843618/tickers', ex)


ex = pickle.dumps(tickers)
qb.ObjectStore.SaveBytes('12843618/tickers', ex)

ticker_list = pickle.loads(bytes(qb.ObjectStore.ReadBytes('12843618/tickers')))

symbols=[]

for i in ticker_list:
    if i not in symbols:
        symbols.append(i)
        
qb.dict_list = {}
i = 0

for x in symbols:
    print(len(symbols)-i)
    i = i+1
    try:
        print(x)
        sym = qb.AddEquity(x).Symbol
        date = qb.GetFundamental(sym, "EarningReports.FileDate", datetime(2015,1,1), datetime.now())
        print(date)
        date = date.drop_duplicates()
        for x in date[x]:
            value = x.to_pydatetime() 
            qb.dict_list.setdefault(sym.Value, []).append(value) 
            qb.dict_list.print

    except KeyError as err:
        symbols.remove(x)
         
    except AttributeError:
        qb.list.setdefault(sym.Value, []).append(x)
    clear_output(wait=True)

Hey Newoptionz,

you just have to use objectsotre if ou want to import tickers from the QC algo to the notebook or from other algos. In this case you have the tickers already in a list which means you can just use this:

qb = QuantBook()
import pickle
tickers = ['PBBF','OXY','MOS', 'AAPL']

i = 0
qb.dict_list = {}
for x in tickers:
    
    try:
        sym = qb.AddEquity(x).Symbol
        date = qb.GetFundamental(sym, "EarningReports.FileDate", datetime(2022,1,1), datetime.now())
        date = date.drop_duplicates()
        for x in date[x]:
            value = x.to_pydatetime() 
            qb.dict_list.setdefault(sym.Value, []).append(value) 
            

    except KeyError as err:
        tickers.remove(x)
        print(f'error: {x}')
        
         
    except AttributeError:
        qb.list.setdefault(sym.Value, []).append(x)
        print(x)

OXY either doesnt seem to have data or the time interval was set when OXY changed name or doesnt trade anymore. I set up try and execpt because sometimes they dont have values. If you want to know you will have to do a manual history request.

Hope it helps ;)

Hi, I see I had a typo on the one stock it should be ‘PBF’.  When I shut down the Juypter test.ipynb and re-open it, now the cells run, but the print statements don't show any output, there is just ‘…’ (three dots) where the output should be, I had a print(symbols) that worked befor and the print(date) worked aswell.  

You either have to restart the kernel. I tried it again and added the print statements that you used. It works with mine… (NOTE: I changed qb to self.  self = qb)

for x in tickers:
    try:
        print(x)
        sym = self.AddEquity(x).Symbol
        date = self.GetFundamental(sym, "EarningReports.FileDate", datetime(2021,1,1), datetime.now())
        print(date)
        date = date.drop_duplicates()
        for x in date[x]:
            value = x.to_pydatetime() 
            self.dict_list.setdefault(sym.Value, []).append(value) 
            self.dict_list.print

    except KeyError as err:
        tickers.remove(x)
         
    except AttributeError:
        self.dict_list.setdefault(sym.Value, []).append(x)

Hi Nico Xenox 

No, still does not work.  Both print(date) work, but when it trys to enter ‘for x in date[x]:’  the code bombs.  I think it trys to convert the dataframes date values to a array and then it tries to add that array to a dictionary using the symbol as the key, but it fails.  Not sure how you would work with such a dictionary if it worked?  That would be a whole other adventure.

The following is not working

for x in date[x]:
            value = x.to_pydatetime() 
            qb.dict_list.setdefault(sym.Value, []).append(value) 

qb = QuantBook()
import pickle
tickers = ['MU','AAPL']

i = 0
qb.dict_list = {}
for x in tickers:
    
    try:
        sym = qb.AddEquity(x).Symbol
        date = qb.GetFundamental(sym, "EarningReports.FileDate", datetime(2022,1,1), datetime.now())
        print(date)
        date = date.drop_duplicates()
        print(date)
        for x in date[x]:
            value = x.to_pydatetime() 
            qb.dict_list.setdefault(sym.Value, []).append(value) 
            

    except KeyError as err:
        tickers.remove(x)
        print(f'error: {x}')
        
         
    except AttributeError:
        qb.list.setdefault(sym.Value, []).append(x)
        print(x)

print(qb.dict_list)

Hey Newoptionz, 

you're right. When executing the code you provided me I get an empty dictionary. But the way I did it it worked. So I tried to find the problem and even tho I found it, I absolutly dont understand why that is so…

You just have to add these three lines to the code(It might be less, didnt test it):

aapl = self.AddEquity("WLL").Symbol
self.AddEquity("BLK").Symbol
history = qb.History(qb.Securities.Keys, 360, Resolution.Daily)

I even checked if the dates of WLL or BLK are the ones you see on the dictionary but that isnt the case. It does work with the ones provided in the tickers list. Here is an image of the result adding these lines to your code:

It seems like that the notebook needs some sort of request to work or something like that I might be wrong tho.

The code(just added the lines above):

qb = QuantBook()
self = qb
aapl = self.AddEquity("WLL").Symbol
self.AddEquity("BLK").Symbol
history = qb.History(qb.Securities.Keys, 360, Resolution.Daily)



tickers = ['MU','AAPL']
qb.dict_list = {}

i = 0

for x in tickers:
    
    try:
        sym = qb.AddEquity(x).Symbol
        date = qb.GetFundamental(sym, "EarningReports.FileDate", datetime(2022,1,1), datetime.now())
        print(date)
        date = date.drop_duplicates()
        print(date)
        for x in date[x]:
            value = x.to_pydatetime() 
            qb.dict_list.setdefault(sym.Value, []).append(value) 
            

    except KeyError as err:
        tickers.remove(x)
        print(f'error: {x}')
        
         
    except AttributeError:
        qb.list.setdefault(sym.Value, []).append(x)
        print(x)

print(qb.dict_list)

It worked for me because every time I started up the notebook I had this block:

import numpy as pd
import pandas as pd
qb = QuantBook()
self = qb
aapl = self.AddEquity("WLL").Symbol
self.AddEquity("BLK").Symbol
history = qb.History(qb.Securities.Keys, 360, Resolution.Daily)
self.dict = {}
from statistics import mean
from statistics import median
from statistics import stdev
dates = datetime(2014,7,23)
import math

Either way, I hope it helps ;)

Nico Xenox 

Thanks, that worked.  I am at the following point, which works, but not sure what to do with the data, other to say it is an interesting excercise in data manipulation

qb = QuantBook()
import pickle
tickers = ['AAPL','MU']

#aapl = qb.AddEquity("AAPL").Symbol
qb.AddEquity("BLK").Symbol
history = qb.History(qb.Securities.Keys, 720, Resolution.Daily)

i = 0
qb.dict_list = {}
for x in tickers:
    
    try:
        sym = qb.AddEquity(x).Symbol
        df = qb.GetFundamental(sym, "EarningReports.FileDate", datetime(2021,2,1), datetime.now())
        df1 = df[sym].unique()
        
        #print(df1)
        
        qb.dict_list.setdefault(sym.Value, df1)
            

    except KeyError as err:
        #tickers.remove(x)
        print(f'error: {x}')
        
         
    except AttributeError:
        #qb.list.setdefault(sym.Value, []).append(x)
        print(f'error: {x}')

print(qb.dict_list)

Hey Newoptionz,

now that you have this data you will be able to do earnings plays for each stock. just save the dictionary into the objectstore and use it in a algorithm. Now the advantage is that you know the date before it will happen, normally you are not able to do that. 

Hi Nico Xenox, yes well that would work for backtesting earnings plays, not really for live trading. I will try give the backtesting a go and post here results etc…