Unlock Stock Insights: An Apple Price Analysis with NumPy

We demonstrate how to read the AAPL.csv file using numpy.loadtxt to extract closing prices and volumes, then compute simple statistics such as arithmetic mean, weighted average (VWAP), maximum, minimum, range, median, and variance.

Basic Statistics

import numpy as np
c, v = np.loadtxt('AAPL.csv', delimiter=',', usecols=(1,2), unpack=True)
mean_c = np.mean(c)
print(mean_c)  # 172.614918033
vwap = np.average(c, weights=v)
print(vwap)   # 170.950010035
print(np.max(c))   # 181.72
print(np.min(c))   # 155.15
print(np.ptp(c))   # 26.57
print(np.median(c))  # 174.35
print(np.var(c))   # 37.5985528621

Daily Returns and Volatility

Daily returns are calculated as the negative difference of consecutive closing prices divided by the previous price, and their standard deviation indicates volatility.

returns = -np.diff(c) / c[1:]
print(np.std(returns))  # 0.0150780328454
positive_days = np.where(returns > 0)
print(positive_days)

Log Returns and Annualized Volatility

Log returns are computed using np.log, and historical volatility is derived from the standard deviation of log returns, scaled by the square root of the number of trading days (252).

logreturns = -np.diff(np.log(c))
volatility = np.std(logreturns) / np.mean(logreturns)
annual_volatility = volatility / np.sqrt(1./252.)
print(volatility)          # 100.096757388
print(annual_volatility)   # 1588.98676256

Date Handling

Since numpy.loadtxt expects numeric data, a custom converter transforms byte‑encoded date strings into weekday numbers.

import datetime
def datestr2num(bytedate):
    return datetime.datetime.strptime(bytedate.decode('utf-8'), '%Y/%m/%d').date().weekday()

dates, c = np.loadtxt('AAPL.csv', delimiter=',', usecols=(0,1), converters={0: datestr2num}, unpack=True)

Weekly Average Closing Prices

Using the weekday indices, we compute the average closing price for each day of the week and identify the highest and lowest averages.

averages = np.zeros(5)
for i in range(5):
    idx = np.where(dates == i)
    prices = np.take(c, idx)
    averages[i] = np.mean(prices)
    print(f"Day {i} avg={averages[i]}")

top = np.max(averages)
top_index = np.argmax(averages)
bot = np.min(averages)
bot_index = np.argmin(averages)
print(f"highest:{top}, top day is {top_index}")
print(f"lowest:{bot}, bottom day is {bot_index}")

Additional useful NumPy functions such as clip for value limiting and compress for conditional filtering are also shown.

a = np.arange(5)
print(a.clip(1,3))   # [1 1 2 3 3]
print(a.compress(a > 2))  # [3 4]

The article emphasizes that NumPy provides a rich set of low‑level functions for data manipulation, serving as a foundation before moving to higher‑level libraries.