Time Series Prediction Interval - Stock Price Predictions Animation Code

6_stock_price_pred.py

### Time Series Prediction Interval (Final Explanation) Stock Price Prediction Example

!pip install --upgrade pandas
!pip install --upgrade pandas-datareader
!pip install celluloid

from celluloid import Camera as Cam
from IPython.display import HTML, clear_output

import pandas_datareader as pdr
import pandas as pd
from statsmodels.tsa.ar_model import AutoReg
from statsmodels.tools.eval_measures import rmse
import numpy as np
import seaborn as sb
import matplotlib.pyplot as plt
from matplotlib import rcParams
from cycler import cycler 

shares_df = pdr.DataReader('AAPL', 'yahoo', start='2021-01-01', end='2021-12-31')
data = shares_df['Close']
data = data.asfreq('d')
data = pd.DataFrame(data.interpolate())

model_fit = AutoReg(data, lags=1).fit()
pred = model_fit.get_prediction(start=pd.to_datetime('2022-01-01'),end = pd.to_datetime('2022-03-31'), dynamic=False)
predictions = pd.DataFrame(pred.predicted_mean)

pred = model_fit.get_prediction(start=pd.to_datetime('2021-07-01'),end = pd.to_datetime('2021-12-31'), dynamic=False)
predictions = pd.DataFrame(pred.predicted_mean)
error = rmse(predictions.predicted_mean, data.Close[-len(predictions):])

pred = model_fit.get_prediction(start=pd.to_datetime('2022-01-01'),end = pd.to_datetime('2022-03-31'), dynamic=False)
predictions = pd.DataFrame(pred.predicted_mean)

predictions['Upper'],predictions['Lower'] = format(0, '.6f'), format(0, '.6f')
for i in range(len(predictions)):
    predictions.Upper[i], predictions.Lower[i] = predictions.predicted_mean[i] + 1.96 * np.sqrt(i+1) * error, predictions.predicted_mean[i] - 1.96 * np.sqrt(i+1) * error

fig, ax = plt.subplots(dpi = 80)
cam = Cam(fig)

rcParams['figure.figsize'] = 17,5
rcParams['axes.spines.top'] = False
rcParams['axes.spines.right'] = False
rcParams['lines.linewidth'] = 2.5

plt.title("Stock Price Prediction")
plt.xlabel("Date")
plt.ylabel("Price (in $)")
sb.despine(right = True, top = True)

sb.lineplot(data = data, x = data.index, y = data.Close, label = 'Historical Data')
plt.axvline(x = predictions.index[0], ymin = 0, ymax = 1, color = 'red', label = 'Prediction Starts from Here')
sb.lineplot(data = predictions, x = predictions.index, y = predictions.predicted_mean, color = '#800000', label = 'Predictions')
sb.lineplot(data = predictions, x = predictions.index, y = predictions.Upper, color = '#00FFE6', label = 'Upper Bound')
sb.lineplot(data = predictions, x = predictions.index, y = predictions.Lower, color = '#38B8EF', label = 'Lower Bound')

fig, ax = plt.subplots(dpi = 400)
cam = Cam(fig)

rcParams['figure.figsize'] = 15,5
rcParams['axes.spines.top'] = False
rcParams['axes.spines.right'] = False
rcParams['lines.linewidth'] = 2.5

plt.title("Stock Price Prediction")
plt.xlabel("Date")
plt.ylabel("Price (in $)")
sb.despine(right = True, top = True)

for i in range(len(data)):
    sb.lineplot(data = data.iloc[0:i], x = data.index[0:i], y = data.Close[0:i], color = '#000080')
    cam.snap()

for i in range(len(predictions)):
    sb.lineplot(data = data, x = data.index, y = data.Close, color = '#000080')
    plt.axvline(x = predictions.index[0], ymin = 0, ymax = 1, color = 'red')
    sb.lineplot(data = predictions.iloc[0:i], x = predictions.index[0:i], y = predictions.predicted_mean[0:i], color = '#800000')
    sb.lineplot(data = predictions.iloc[0:i], x = predictions.index[0:i], y = predictions.Upper[0:i], color = '#00FFE6')
    sb.lineplot(data = predictions.iloc[0:i], x = predictions.index[0:i], y = predictions.Lower[0:i], color = '#38B8EF')
    cam.snap()

for i in range(1,len(data)):
    sb.lineplot(data = data, x = data.index, y = data.Close, color = '#000080')
    plt.axvline(x = predictions.index[0], ymin = 0, ymax = 1, color = 'red')
    sb.lineplot(data = predictions, x = predictions.index, y = predictions.predicted_mean, color = '#800000')
    sb.lineplot(data = predictions, x = predictions.index, y = predictions.Upper, color = '#00FFE6')
    sb.lineplot(data = predictions, x = predictions.index, y = predictions.Lower, color = '#38B8EF')
    cam.snap()
plt.close(fig)


animation = cam.animate(blit=False, interval=15)
HTML(animation.to_html5_video())