Visualizing Stock Data and Building K‑Line Charts with Python

Data Import and Cleaning

Stock data is stored in stockData.txt. It is read into a pandas DataFrame with

pd.read_table('stockData.txt', usecols=range(15), parse_dates=[0], index_col=0)

. The column name ' open' contains leading spaces and is renamed to 'open' using stock.rename(columns={' open':'open'}, inplace=True). The first five rows are displayed with stock.head() and the DataFrame information is inspected via stock.info().

Data Observation

The dataset includes price‑related indicators (open, close, high, low, price_change, ma5, ma10, ma20) and volume‑related indicators (volume, turnover, v_ma5, v_ma10, v_ma20). A table image illustrates these metrics.

Time Series Plot

The closing price is plotted over time with stock['close'].plot(grid=True), producing a line chart that shows price fluctuations.

K‑Line (Candlestick) Chart

Matplotlib’s candlestick_ohlc function from the matplotlib.finance module is used to draw a candlestick chart. A helper function pandas_candlestick_ohlc(stock_data, otherseries=None) sets up the figure, formats the date axis, and calls candlestick_ohlc with appropriate colors. The resulting chart shows daily open, high, low, and close prices, where red indicates an up day and green a down day.

Relative Change

Relative price change is calculated by dividing the closing price by the initial price: stock['return'] = stock['close'] / stock['close'].iloc[0]. The return series is plotted with a zero reference line.

Correlation Analysis

A subset of columns ( close, price_change, ma20, volume, v_ma20, turnover) is selected and visualized with pd.scatter_matrix to examine pairwise relationships. A correlation matrix is computed with np.corrcoef(small.T) and displayed using plt.matshow. The matrix reveals a strong positive correlation (≈0.91) between closing price and volume.

Dual‑Axis Plot of Close and Volume

The closing price and volume are plotted together with separate y‑axes using

stock[['close','volume']].plot(secondary_y='volume', grid=True)

.

Moving Average Strategy

Google stock data is fetched from Yahoo using pandas_datareader. After renaming columns to lowercase, 5‑day and 20‑day moving averages are computed with

stock['ma5'] = np.round(stock['close'].rolling(window=5, center=False).mean(), 2)

and

stock['ma20'] = np.round(stock['close'].rolling(window=20, center=False).mean(), 2)

. The difference stock['ma5-20'] = stock['ma5'] - stock['ma20'] is signed and plotted to identify crossover points. A signal series is derived by taking the sign of the difference’s first derivative:

stock['signal'] = np.sign(stock['diff'] - stock['diff'].shift(1))

. Buy signals (1) and sell signals (‑1) are plotted.

Back‑Testing the Strategy

Buy and sell dates are extracted from the signal series and combined into a trade DataFrame that records the operation (Buy/Sell) and the closing price on that day. Sorting the trades shows two complete cycles, but both sell prices are lower than the corresponding buy prices, resulting in a loss. The tutorial notes that this simple moving‑average crossover is only illustrative and not a guaranteed profitable strategy.

The article concludes that while the moving‑average crossover can be useful over longer time horizons, real‑world trading requires more sophisticated risk management and portfolio allocation.