I'm new to python, but am falling in love with it as my understanding gets better. That being said I've managed to piece together a script to dump the Quandl WIKI EOD database into CSV's for each respective stock. which I've included below (Python 2.7):
import pandas as pd
import Quandl
#this ignores the first symbol in tickers.txt, add a header!
symbols_list = list(pd.read_csv('tickers.txt', index_col=0).T)
print symbols_list
df = {}
for ticker in symbols_list:
stock = str("WIKI/" + ticker + ".11")
#print stock
print str(ticker + ".csv")
df[ticker] = Quandl.get(stock, authtoken="XXXXXXXXXXXXX", trim_start="2016-01-01")
df[ticker].to_csv(str(ticker + ".csv"))
My next goal is to create a historical running high and low column for each stock csv. This would make it easy to create a historical "gas gauge" for the stock market similar to the ones on http://www.finviz.com/. The new high/new low gauge is extremely useful. When 2000+ stocks make a new low and only a few make a new high, it's easy to gauge just how hard a sell off is.
So far I've managed to iterate through each .csv in my working folder and keep a tally of each new high / new low. My hang up is on the proper way to enter this information into each csv as new columns. The EOD CSV's have ['Date', 'Adj. Close'] for headers in each, and am I looking to add ['High', 'Low'].
My code so far:
import os
import pandas as pd
csv = [f for f in os.listdir('.') if f.endswith('.csv')]
print csv
for i in csv:
df = pd.read_csv(i, index_col=0, parse_dates=True)
print df
j=0
k=0
for i in df['Adj. Close']: #for each value in adj close
if j==0:
j=i
k=i
elif i>j:
j=i #high
elif i<k:
k=i #low
print j,k #print high & low
I really appreciate your help guys! All of the code snippets and answered questions have made learning python so much more enjoyable.
Here's the output of the script so far:
['A.csv', 'AA.csv']
Adj. Close
Date
2016-01-04 40.69
2016-01-05 40.55
2016-01-06 40.73
2016-01-07 39.00
2016-01-08 38.59
2016-01-11 37.94
2016-01-12 38.19
2016-01-13 36.86
2016-01-14 37.61
2016-01-15 37.11
2016-01-19 37.20
2016-01-20 37.26
2016-01-21 37.23
2016-01-22 37.98
40.69 40.69
40.69 40.55
40.73 40.55
40.73 39.0
40.73 38.59
40.73 37.94
40.73 37.94
40.73 36.86
40.73 36.86
40.73 36.86
40.73 36.86
40.73 36.86
40.73 36.86
40.73 36.86
Adj. Close
Date
2016-01-04 9.71
2016-01-05 9.27
2016-01-06 8.61
2016-01-07 8.27
2016-01-08 8.07
2016-01-11 8.00
2016-01-12 7.28
2016-01-13 7.13
2016-01-14 7.24
2016-01-15 6.90
2016-01-19 6.74
2016-01-20 6.74
2016-01-21 7.09
2016-01-22 6.87
9.71 9.71
9.71 9.27
9.71 8.61
9.71 8.27
9.71 8.07
9.71 8.0
9.71 7.28
9.71 7.13
9.71 7.13
9.71 6.9
9.71 6.74
9.71 6.74
9.71 6.74
9.71 6.74
This code worked beautifully:
import os
import pandas as pd
csv = [f for f in os.listdir('.') if f.endswith('.csv')]
print csv
for i in csv:
df = pd.read_csv(i, index_col=0, parse_dates=True)
#print df
df['High'] = pd.rolling_max(df['Adj. Close'], window=df.shape[0], min_periods=1)
df['Low'] = pd.rolling_min(df['Adj. Close'], window=df.shape[0], min_periods=1)
#print df
g = os.path.normcase(os.path.join(os.path.dirname(__file__), i)) #os.getcwd()
print 'full path =', g
with open(g , "w") as f:
df.to_csv(f)
Now I'll see about creating a historical list of each day's new high/new low
You can use length of df - shape as window of rolling_min and rolling_max:
print df
Date Adj. Close
0 2016-01-04 9.71
1 2016-01-05 9.27
2 2016-01-06 8.61
3 2016-01-07 8.27
4 2016-01-08 8.07
5 2016-01-11 8.00
6 2016-01-12 7.28
7 2016-01-13 7.13
8 2016-01-14 7.24
9 2016-01-15 6.90
10 2016-01-19 6.74
11 2016-01-20 6.74
12 2016-01-21 7.09
13 2016-01-22 6.87
df['high'] = pd.rolling_max(df['Adj. Close'], window=df.shape[0], min_periods=1)
df['low'] = pd.rolling_min(df['Adj. Close'], window=df.shape[0], min_periods=1)
print df
Date Adj. Close high low
0 2016-01-04 9.71 9.71 9.71
1 2016-01-05 9.27 9.71 9.27
2 2016-01-06 8.61 9.71 8.61
3 2016-01-07 8.27 9.71 8.27
4 2016-01-08 8.07 9.71 8.07
5 2016-01-11 8.00 9.71 8.00
6 2016-01-12 7.28 9.71 7.28
7 2016-01-13 7.13 9.71 7.13
8 2016-01-14 7.24 9.71 7.13
9 2016-01-15 6.90 9.71 6.90
10 2016-01-19 6.74 9.71 6.74
11 2016-01-20 6.74 9.71 6.74
12 2016-01-21 7.09 9.71 6.74
13 2016-01-22 6.87 9.71 6.74