Main study : follow-along practice with the book, Python Securities Data Analysis <- 65%
파이썬 증권 데이터 분석
웹 스크레이핑으로 증권 데이터를 주기적으로 자동 수집, 분석, 자동 매매, 예측하는 전 과정을 파이썬으로 직접 구현한다. 그 과정에서 금융 데이터 처리 기본 라이브러리(팬더스)부터 주가 예
www.aladin.co.kr
Before, when I posted about coding studies, I tried to make them tutorial-like.
Looking back now, that's not great.
The problem isn't that there isn't enough information — it's that there's too much, isn't it?
So I'll settle for just jotting down the main keywords and links.
Setup
64-bit Python (Windows x86-64 executable installer) download -> install Python (*check the option to auto-add to PATH)
-> from cmd install packages (pip install matplotlib, and pythonexe -m pip install --upgrade pip) *handling EnvironmentError
-> install packages for practice (save the list as a file and run it from cmd = pip install -r filename.txt)
-> 32-bit Python (Windows x86 executable installer) download -> install (*don't check the PATH option)
Main references
Sub references
- PyQt GUI for Stock Investing
- Securities Data Analysis (hands-on review + video)
Main keywords
pip install .. runs in the terminal, not in Python
empty list, dict, tuple, set
- #list ls=[], #dict d={}, #tuple t=(), #set s=set()
List
- append(), extend()
comprehension
dictionary
- a.key(), a.values(), a.items(), a.clear(), a.get('key'), a.get('key','default comment')
- f-string, {}, $s
set
- no duplicates, no indexing, supports intersection/difference/union operations
- much faster "lookup (if)" speed (for-loops are a bit slow) *timeit (speed test LB)
Library = modules (.py files) and packages (folders)
from module/datetime import method/datetime as alias/dt
print(dt.now())
myPackage.moduleA.functionA()
package.module.function
objects, instances, inheritance
requests python -m pip install requests
with ~ as file-object
SHA-256 import hashlib
Basic math
mean, variance, standard deviation
linear regression model (cause x, continuous outcome y, measurement error e, expressing their relationship as a linear equation (line) :
stats.linregress(df['target1'],df['target2']) )
variance, covariance, correlation coefficient, correlation
Key indicator functions
CAGR Compound Annual Growth Rate
def getGAGR(first,last,years): # compound annual growth / yearly compounded return (sales volume, user growth rate)
return (last/first)**(1/years)-1
R1 Daily percent change daily rate of change (comparing two stocks at different prices)
R1 (today's rate of change) = (target.['column'].shift(1) - target.['column'] / target.['column'].shift(1) ) * 100
MDD Maximum Drawdown maximum loss drawdown
(low - high) / low
Sharpe Ratio
(expected portfolio return - risk-free rate) / standard deviation of returns
Bollinger Bands + related practice reference (Class101 — TimePercent)
upper Bollinger Band ubb = middle Bollinger Band + (2 × stdev)
middle Bollinger Band mbb = 20-day moving average of close
lower Bollinger Band lbb = middle Bollinger Band - (2 × stdev)
PerB (Price, Length, Mult, MaKind, [Optional]Pos) = %b *additional reference
= (price – lower) / (upper – lower) = (close - lbb) / (ubb - lbb)
* Price: the base price data for the moving average (open, high, low, close)
* Length: moving-average period
* Mult: multiplier
* MaKind: method for computing the moving average
* (S: simple MA, E: exponential MA, W: weighted MA, A: cumulative MA)
Bandwidth
= (upper – lower) / middle = (ubb - lbb) / mbb
MFI Money Flow Index cash-flow indicator
- uses typical price (not close) and volume (which leads price) as an indicator
RSI relative strength index
- RS = average upswing over n days / average downswing over n days
- RSI = 100 - 100 / (1 + RS)
API, Library
Yahoo Finance
Korea Corporate Disclosure Channel
Naver Finance *handling crawler blocks
backtrader : bt.indicators.(reference)
Data analytics libraries
numpy numerical python | cmd > pip install numpy | examples
pandas | cmd > pip install pandas | describe(), Series, DataFrame (multiple Series combined under one index), regression analysis
SciPy (pronounced “Sigh Pie”) is a Python-based ecosystem of open-source software for mathematics, science, and engineering. In particular, these are some of the core packages | cmd > pip install scipy
html5lib parsing HTML | cmd >pip install html5lib
beautifulsoup4 a Python library for pulling data out of HTML and XML files | cmd >pip install beautifulsoup4 | example 1 example 2
Visualization
matplotlib | cmd > pip install matplotlib | drawing images, drawing charts
plotly (& example: showing plot, plotly, and dash together on one page, plotly libraries )
front-end framework
Django - (Vitor Freitas's blog)
