Contents

Apple Stock

Apple Stock

Author: Mia [Nguyen Thao My] Ta

Course Project, UC Irvine, Math 10, S22

Introduction

The purpose of the project is to estimate Apple’s future stock price using historical data and linear regression. Analyzing stock data to predict the expected value of a company’s stock and displaying stock movement over time.

Section 1: Clean up Dataset

import pandas as pd
import numpy as np
import altair as alt
from altair import Chart, X, Y
import matplotlib.pyplot as plt
import seaborn as sns

df = pd.read_csv("AAPL.csv")
df = df.dropna()
df
Unnamed: 0 symbol date close high low open volume adjClose adjHigh adjLow adjOpen adjVolume divCash splitFactor
0 0 AAPL 2015-05-27 00:00:00+00:00 132.045 132.260 130.0500 130.34 45833246 121.682558 121.880685 119.844118 120.111360 45833246 0.0 1.0
1 1 AAPL 2015-05-28 00:00:00+00:00 131.780 131.950 131.1000 131.86 30733309 121.438354 121.595013 120.811718 121.512076 30733309 0.0 1.0
2 2 AAPL 2015-05-29 00:00:00+00:00 130.280 131.450 129.9000 131.23 50884452 120.056069 121.134251 119.705890 120.931516 50884452 0.0 1.0
3 3 AAPL 2015-06-01 00:00:00+00:00 130.535 131.390 130.0500 131.20 32112797 120.291057 121.078960 119.844118 120.903870 32112797 0.0 1.0
4 4 AAPL 2015-06-02 00:00:00+00:00 129.960 130.655 129.3200 129.86 33667627 119.761181 120.401640 119.171406 119.669029 33667627 0.0 1.0
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
1253 1253 AAPL 2020-05-18 00:00:00+00:00 314.960 316.500 310.3241 313.17 33843125 314.960000 316.500000 310.324100 313.170000 33843125 0.0 1.0
1254 1254 AAPL 2020-05-19 00:00:00+00:00 313.140 318.520 313.0100 315.03 25432385 313.140000 318.520000 313.010000 315.030000 25432385 0.0 1.0
1255 1255 AAPL 2020-05-20 00:00:00+00:00 319.230 319.520 316.2000 316.68 27876215 319.230000 319.520000 316.200000 316.680000 27876215 0.0 1.0
1256 1256 AAPL 2020-05-21 00:00:00+00:00 316.850 320.890 315.8700 318.66 25672211 316.850000 320.890000 315.870000 318.660000 25672211 0.0 1.0
1257 1257 AAPL 2020-05-22 00:00:00+00:00 318.890 319.230 315.3500 315.77 20450754 318.890000 319.230000 315.350000 315.770000 20450754 0.0 1.0

1258 rows × 15 columns

Define DataFrame and getting exact weekday and month.

df = df.iloc[:,1:15]
df["date"] = pd.to_datetime(df["date"])
df = df.sort_values(by="date")
df["weekday"] = df["date"].dt.day_name()
df["month"] = df["date"].dt.month
df
symbol date close high low open volume adjClose adjHigh adjLow adjOpen adjVolume divCash splitFactor weekday month
0 AAPL 2015-05-27 00:00:00+00:00 132.045 132.260 130.0500 130.34 45833246 121.682558 121.880685 119.844118 120.111360 45833246 0.0 1.0 Wednesday 5
1 AAPL 2015-05-28 00:00:00+00:00 131.780 131.950 131.1000 131.86 30733309 121.438354 121.595013 120.811718 121.512076 30733309 0.0 1.0 Thursday 5
2 AAPL 2015-05-29 00:00:00+00:00 130.280 131.450 129.9000 131.23 50884452 120.056069 121.134251 119.705890 120.931516 50884452 0.0 1.0 Friday 5
3 AAPL 2015-06-01 00:00:00+00:00 130.535 131.390 130.0500 131.20 32112797 120.291057 121.078960 119.844118 120.903870 32112797 0.0 1.0 Monday 6
4 AAPL 2015-06-02 00:00:00+00:00 129.960 130.655 129.3200 129.86 33667627 119.761181 120.401640 119.171406 119.669029 33667627 0.0 1.0 Tuesday 6
... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ... ...
1253 AAPL 2020-05-18 00:00:00+00:00 314.960 316.500 310.3241 313.17 33843125 314.960000 316.500000 310.324100 313.170000 33843125 0.0 1.0 Monday 5
1254 AAPL 2020-05-19 00:00:00+00:00 313.140 318.520 313.0100 315.03 25432385 313.140000 318.520000 313.010000 315.030000 25432385 0.0 1.0 Tuesday 5
1255 AAPL 2020-05-20 00:00:00+00:00 319.230 319.520 316.2000 316.68 27876215 319.230000 319.520000 316.200000 316.680000 27876215 0.0 1.0 Wednesday 5
1256 AAPL 2020-05-21 00:00:00+00:00 316.850 320.890 315.8700 318.66 25672211 316.850000 320.890000 315.870000 318.660000 25672211 0.0 1.0 Thursday 5
1257 AAPL 2020-05-22 00:00:00+00:00 318.890 319.230 315.3500 315.77 20450754 318.890000 319.230000 315.350000 315.770000 20450754 0.0 1.0 Friday 5

1258 rows × 16 columns

We can observe the distribution of data points on different days of the week from the chart displayed below.

days = [pd.to_datetime(f"2022-06-{x}").day_name() for x in range (6,13,1)]
df[days] = 0
for i in days:
    df.loc[df["weekday"] == i,i] = 1
df[days].sum(axis=0)

Monday       236
Tuesday      257
Wednesday    257
Thursday     255
Friday       253
Saturday       0
Sunday         0
dtype: int64

Section 2: Data Graph

Let’s take a look at the Apple stock trends for the past couple years.

mark = alt.selection_single(nearest = True, on = "mouseover", fields = ["date"], empty = 'none')

main = alt.Chart().mark_point().encode(
    x="date:T",
    tooltip=["date", "weekday", "adjClose", "close"],
    opacity=alt.value(0),
).add_selection(mark)
line = alt.Chart().mark_line().encode(
        x= alt.X("date", title="Date"),
        y= alt.Y("adjClose",  axis = alt.Axis(title="", format = "$f")),
    )

rule = alt.Chart().mark_rule(color="gray").encode(
    x="date:T",
).transform_filter(mark)

c = alt.layer(main,line,rule, data = df, title = "Apple Stock Price").add_selection(mark)
c

In stead of looking at all the data points, we could gather a subdata and look at the trend at that specific period. The following function day(x) will take an input x and displays a DataFrame of all data points with the same day x. In this case, we are looking at the DataFrame on the 5th of every month. If there is a missing month, for example, we go from June 5th, 2015 to August 5th, 2015, then July 5th, 2017 must fall on the weekend (either Saturday or Sunday).

#This is a subDataFrame of all 5th days.
def day(x):
    return df[df["date"].dt.day == x]
day(5)
symbol date close high low open volume adjClose adjHigh adjLow ... splitFactor weekday month Monday Tuesday Wednesday Thursday Friday Saturday Sunday
7 AAPL 2015-06-05 00:00:00+00:00 128.65 129.6900 128.3600 129.500 35626800 118.553986 119.512370 118.286744 ... 1.0 Friday 6 0 0 0 0 1 0 0
49 AAPL 2015-08-05 00:00:00+00:00 115.40 117.4400 112.1000 112.950 98384461 106.343801 108.223708 103.302773 ... 1.0 Wednesday 8 0 0 1 0 0 0 0
91 AAPL 2015-10-05 00:00:00+00:00 110.78 111.3698 109.0700 109.880 52064743 102.547449 103.093418 100.964527 ... 1.0 Monday 10 1 0 0 0 0 0 0
114 AAPL 2015-11-05 00:00:00+00:00 120.92 122.6900 120.1800 121.850 39552680 112.415257 114.060767 111.727304 ... 1.0 Thursday 11 0 0 0 1 0 0 0
154 AAPL 2016-01-05 00:00:00+00:00 102.71 105.8500 102.4100 105.750 55790992 95.486033 98.405185 95.207133 ... 1.0 Tuesday 1 0 1 0 0 0 0 0
176 AAPL 2016-02-05 00:00:00+00:00 94.02 96.9200 93.6900 96.520 46418064 87.877747 90.588292 87.569306 ... 1.0 Friday 2 0 0 0 0 1 0 0
216 AAPL 2016-04-05 00:00:00+00:00 109.81 110.7300 109.4200 109.510 26578652 102.636199 103.496096 102.271677 ... 1.0 Tuesday 4 0 1 0 0 0 0 0
238 AAPL 2016-05-05 00:00:00+00:00 93.24 94.0700 92.6800 94.000 35890500 87.681466 88.461986 87.154851 ... 1.0 Thursday 5 0 0 0 1 0 0 0
279 AAPL 2016-07-05 00:00:00+00:00 94.99 95.4000 94.4600 95.390 27705210 89.327139 89.712697 88.828736 ... 1.0 Tuesday 7 0 1 0 0 0 0 0
302 AAPL 2016-08-05 00:00:00+00:00 107.48 107.6500 106.1800 106.270 40553402 101.616715 101.777441 100.387633 ... 1.0 Friday 8 0 0 0 0 1 0 0
344 AAPL 2016-10-05 00:00:00+00:00 113.05 113.6600 112.6900 113.400 21453089 106.882858 107.459581 106.542497 ... 1.0 Wednesday 10 0 0 1 0 0 0 0
386 AAPL 2016-12-05 00:00:00+00:00 109.11 110.0300 108.2500 110.000 34324540 103.693167 104.567493 102.875862 ... 1.0 Monday 12 1 0 0 0 0 0 0
407 AAPL 2017-01-05 00:00:00+00:00 116.61 116.8642 115.8100 115.920 22193587 110.820825 111.062405 110.060541 ... 1.0 Thursday 1 0 0 0 1 0 0 0
469 AAPL 2017-04-05 00:00:00+00:00 144.02 145.4600 143.8100 144.220 27717854 137.459193 138.833594 137.258760 ... 1.0 Wednesday 4 0 0 1 0 0 0 0
490 AAPL 2017-05-05 00:00:00+00:00 148.96 148.9800 146.7600 146.760 26787359 142.174152 142.193241 140.074373 ... 1.0 Friday 5 0 0 0 0 1 0 0
510 AAPL 2017-06-05 00:00:00+00:00 153.93 154.4500 153.4600 154.340 24803858 147.518967 148.017310 147.068542 ... 1.0 Monday 6 1 0 0 0 0 0 0
531 AAPL 2017-07-05 00:00:00+00:00 144.09 144.7900 142.7237 143.690 20758795 138.088794 138.759639 136.779399 ... 1.0 Wednesday 7 0 0 1 0 0 0 0
574 AAPL 2017-09-05 00:00:00+00:00 162.08 164.2500 160.5600 163.750 29317054 155.959566 158.047623 154.496964 ... 1.0 Tuesday 9 0 1 0 0 0 0 0
596 AAPL 2017-10-05 00:00:00+00:00 155.39 155.4400 154.0500 154.180 21032800 149.522193 149.570305 148.232794 ... 1.0 Thursday 10 0 0 0 1 0 0 0
638 AAPL 2017-12-05 00:00:00+00:00 169.64 171.5200 168.4000 169.060 27008428 163.822840 165.638373 162.625361 ... 1.0 Tuesday 12 0 1 0 0 0 0 0
659 AAPL 2018-01-05 00:00:00+00:00 175.00 175.3700 173.0500 173.440 23016177 168.999039 169.356352 167.115907 ... 1.0 Friday 1 0 0 0 0 1 0 0
679 AAPL 2018-02-05 00:00:00+00:00 156.49 163.8800 156.0000 159.100 66090446 151.123769 158.260357 150.650572 ... 1.0 Monday 2 1 0 0 0 0 0 0
698 AAPL 2018-03-05 00:00:00+00:00 176.82 177.7400 174.5200 175.210 28401366 171.444416 172.336446 169.214339 ... 1.0 Monday 3 1 0 0 0 0 0 0
720 AAPL 2018-04-05 00:00:00+00:00 172.80 174.2304 172.0800 172.580 26933197 167.546630 168.933543 166.848519 ... 1.0 Thursday 4 0 0 0 1 0 0 0
762 AAPL 2018-06-05 00:00:00+00:00 193.31 193.9400 192.3600 193.065 21565963 188.158618 188.771829 187.233934 ... 1.0 Tuesday 6 0 1 0 0 0 0 0
783 AAPL 2018-07-05 00:00:00+00:00 185.40 186.4100 184.2800 185.260 16604248 180.459406 181.442491 179.369252 ... 1.0 Thursday 7 0 0 0 1 0 0 0
826 AAPL 2018-09-05 00:00:00+00:00 226.87 229.6700 225.1000 228.990 33332960 221.601064 224.336036 219.872172 ... 1.0 Wednesday 9 0 0 1 0 0 0 0
848 AAPL 2018-10-05 00:00:00+00:00 224.29 228.4100 220.5800 227.960 33580463 219.080983 223.105299 215.457146 ... 1.0 Friday 10 0 0 0 0 1 0 0
869 AAPL 2018-11-05 00:00:00+00:00 201.59 204.3900 198.1700 204.300 66163669 196.908179 199.643150 193.567607 ... 1.0 Monday 11 1 0 0 0 0 0 0
930 AAPL 2019-02-05 00:00:00+00:00 174.18 175.0800 172.3501 172.860 36101628 170.730466 171.612642 168.936806 ... 1.0 Tuesday 2 0 1 0 0 0 0 0
949 AAPL 2019-03-05 00:00:00+00:00 175.53 176.0000 174.5400 175.940 19737419 172.790771 173.253437 171.816221 ... 1.0 Tuesday 3 0 1 0 0 0 0 0
972 AAPL 2019-04-05 00:00:00+00:00 197.00 197.1000 195.9300 196.450 18526644 193.925722 194.024161 192.872420 ... 1.0 Friday 4 0 0 0 0 1 0 0
1013 AAPL 2019-06-05 00:00:00+00:00 182.54 184.9900 181.1400 184.280 29773427 180.393083 182.814268 179.009549 ... 1.0 Wednesday 6 0 0 1 0 0 0 0
1034 AAPL 2019-07-05 00:00:00+00:00 204.23 205.0800 202.9000 203.350 17265518 201.827979 202.667982 200.513622 ... 1.0 Friday 7 0 0 0 0 1 0 0
1055 AAPL 2019-08-05 00:00:00+00:00 193.34 198.6490 192.5800 197.990 52392969 191.066060 196.312619 190.314999 ... 1.0 Monday 8 1 0 0 0 0 0 0
1077 AAPL 2019-09-05 00:00:00+00:00 213.28 213.9700 211.5100 212.000 23946984 211.579012 212.263509 209.823129 ... 1.0 Thursday 9 0 0 0 1 0 0 0
1120 AAPL 2019-11-05 00:00:00+00:00 257.13 258.1900 256.3200 257.050 19974427 255.079292 256.130838 254.275752 ... 1.0 Tuesday 11 0 1 0 0 0 0 0
1141 AAPL 2019-12-05 00:00:00+00:00 265.58 265.8900 262.7300 263.790 18661343 264.243867 264.552308 261.408205 ... 1.0 Thursday 12 0 0 0 1 0 0 0
1182 AAPL 2020-02-05 00:00:00+00:00 321.45 324.7600 318.9500 323.520 29706718 319.832785 323.126133 317.345363 ... 1.0 Wednesday 2 0 0 1 0 0 0 0
1202 AAPL 2020-03-05 00:00:00+00:00 292.92 299.5500 291.4100 295.520 46893219 292.147547 298.760063 290.641529 ... 1.0 Thursday 3 0 0 0 1 0 0 0
1244 AAPL 2020-05-05 00:00:00+00:00 297.56 301.0000 294.4600 295.060 36937795 296.775311 300.206239 293.683485 ... 1.0 Tuesday 5 0 1 0 0 0 0 0

41 rows × 23 columns

sel = alt.selection_single(fields=["month"])
def draw_by_day(x):
    day(x)
    c = alt.Chart(day(x), title = f"Apple Stock Price on Day {x}").mark_circle(color = "red").encode(
        x= alt.X("date", title = "Date"),
        y= alt.Y("adjClose", axis = alt.Axis(title="", format = "$f")),
        tooltip=["date", "weekday", "adjClose", "close"],
        size=alt.condition(sel, alt.value(80),alt.value(30))
    ).add_selection(sel)
    return c

draw_by_day(5)

c+draw_by_day(5)

With function day(x), I got a DataFrame with all 5th days, though when I’m trying to sketch Apple’s stock using this DataFrame, all of the points got shifted to a day prior. I hypothesize that this issue might occur due to the time zone difference since all the dates are 00:00. To confirm this idea, I created column df["Date"] = df["date"].dt.date which contains dates without hour. However, when I could not compute altChart with data from Date, the error alert TypeError: Object of type date is not JSON serializable. I also try to create an alt.Chart with data from Date2, a column containing only days, the output is still the same. Hence, I believe this is Deepnote error.

df["Date2"] = df["date"].dt.day
df["Date2"]

0       27
1       28
2       29
3        1
4        2
        ..
1253    18
1254    19
1255    20
1256    21
1257    22
Name: Date2, Length: 1258, dtype: int64

def day2(x):
    return df[df["Date2"] == x]

def draw_by_day2(x):
    day2(x)
    c = alt.Chart(day(x), title = f"Apple Stock Price on Day {x}").mark_circle(color = "red").encode(
        x= alt.X("date", title = "Date"),
        y= alt.Y("adjClose", axis = alt.Axis(title="", format = "$f")),
        tooltip=["date", "weekday", "adjClose", "close"],
        size=alt.condition(sel, alt.value(80),alt.value(30))
    ).add_selection(sel)
    return c

draw_by_day2(5)

The distribution chart of volume.

sns.distplot(df.volume, bins=25, color="b")
plt.title("Distribution of Apple Volume", fontsize=20, color = 'Purple')
plt.show()

/shared-libs/python3.7/py/lib/python3.7/site-packages/seaborn/distributions.py:2619: FutureWarning: `distplot` is a deprecated function and will be removed in a future version. Please adapt your code to use either `displot` (a figure-level function with similar flexibility) or `histplot` (an axes-level function for histograms).
  warnings.warn(msg, FutureWarning)
../../_images/MiaTa_20_1.png

Section 3: Linear Regression

There are some commonly used indicators such as moving averages (SMA, EMA, MACD), relative Strength index (RSI), or bollinger bands (BBANDS). For our data, we will use exponetial moving average (EMA). First we need to calculate the 15-day Exponential Moving Averages (EMA15) and add EMA15 column to our DataFrame.

df["EMA15"] = df["adjClose"].ewm(span=15, min_periods=0, ignore_na = False, adjust=False).mean()
df_sub = df[["date","adjClose","EMA15"]]
df_sub.info()

<class 'pandas.core.frame.DataFrame'>
Int64Index: 1258 entries, 0 to 1257
Data columns (total 3 columns):
 #   Column    Non-Null Count  Dtype              
---  ------    --------------  -----              
 0   date      1258 non-null   datetime64[ns, UTC]
 1   adjClose  1258 non-null   float64            
 2   EMA15     1258 non-null   float64            
dtypes: datetime64[ns, UTC](1), float64(2)
memory usage: 39.3 KB

plt.plot(df_sub["date"],df["adjClose"],label = "adjClose")
plt.plot(df_sub["date"],df["EMA15"], label = "EMA15")
plt.legend()
plt.show()
../../_images/MiaTa_24_0.png 
chartClose = alt.Chart(df_sub).mark_line(color = "orange").encode(
    x = "date",
    y="adjClose"
)

chartEMA = alt.Chart(df_sub).mark_line(color ="Blue").encode(
    x = "date",
    y = "EMA15"
)

chartClose + chartEMA

Both graphs above are the same; however, I prefer to use plt.plot. When we have multiple linear lines, plt.plot is easier to label different lines. There is a way to create a title for various linear lines, but it is more complicated.

from sklearn.preprocessing import StandardScaler
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.metrics import log_loss, mean_squared_error, mean_absolute_error, r2_score

X_train, X_test, y_train, y_test = train_test_split(df[['adjClose']],df[['EMA15']], test_size=0.2)
reg = LinearRegression()
reg.fit(X_train, y_train)
y_pred = reg.predict(X_test)
print("Mean Absolute Error is", mean_absolute_error(y_test, y_pred))
print("Coefficient of Determination is", r2_score(y_test,y_pred))

Mean Absolute Error is 3.8818754599384966
Coefficient of Determination is 0.9899721051486294

The graph below represents the observed values and predicted values on day x. The red dots are observed values whereas the purple line represents the predicted values.

df["Pred"] = reg.predict(df[["adjClose"]])
def pred_chart(x):
    day(x)
    c = alt.Chart(day(x), title = f"Apple's Stock Predicted and Observed Values on Day {x}").mark_circle(color = "red").encode(
        x= alt.X("date", scale=alt.Scale(zero=False)),
        y= alt.Y("adjClose", scale=alt.Scale(zero=False)),
    )
    c1 = alt.Chart(day(x)).mark_line(color = "Purple"). encode(
        x = alt.X("date",scale=alt.Scale(zero=False)),
        y = alt.Y("Pred",scale=alt.Scale (zero=False))
    )
    return c + c1

pred_chart(5)

Section 4: K-Nearest Neighbors Regressor

from sklearn.neighbors import KNeighborsClassifier, KNeighborsRegressor

K-Nearest Neighbors R egressor is a non-parametric method to estimate the association between the independent and continuous outcomes by averaging observations from the same neighborhood. In this case, I’ll use ‘n neighbors = 10’, which will take the ten closest points to take the average.

X_train, X_test, y_train, y_test = train_test_split(df[['adjClose']],df[['EMA15']], test_size=0.2)
reg1 = KNeighborsRegressor(n_neighbors=10)
reg1.fit(X_train,y_train)
reg1.predict(X_test)
print("Mean Test Absolute Error is", mean_absolute_error(reg1.predict(X_test), y_test))
print("Mean Train Absolute Error is", mean_absolute_error(reg1.predict(X_train), y_train))

Mean Test Absolute Error is 3.6018819215316356
Mean Train Absolute Error is 3.3270751521487245

Function get_scores(k) will take an input k to calculate the mean absolute train error and mean absolute test error and from there I can create a DataFrame of train error and test error for range k from 1 to 150.

def get_scores(k):
    reg = KNeighborsRegressor(n_neighbors=k)
    reg.fit(X_train, y_train)
    train_error = mean_absolute_error(reg.predict(X_train), y_train)
    test_error = mean_absolute_error(reg.predict(X_test), y_test)
    return (test_error, train_error)

get_scores(10)

(3.6018819215316356, 3.3270751521487245)

df_scores = pd.DataFrame({"k":range(1,150),"train_error":np.nan,"test_error":np.nan})
for i in df_scores.index:
    df_scores.loc[i,["train_error","test_error"]] = get_scores(df_scores.loc[i,"k"])

df_scores
k train_error test_error
0 1 4.934897 0.043639
1 2 4.469894 2.293783
2 3 4.152767 2.824419
3 4 3.871697 3.106926
4 5 3.831669 3.206610
... ... ... ...
144 145 5.274543 5.751847
145 146 5.303177 5.779944
146 147 5.327647 5.816586
147 148 5.353606 5.850652
148 149 5.389431 5.880698

149 rows × 3 columns

The blue curve denotes training error, while the orange curve denotes test error. Furthermore, we can see that extremely high K values cause underfitting, whereas low K values cause overfitting to occur.

df_scores["kinv"] = 1/df_scores.k
ctrain = alt.Chart(df_scores).mark_line().encode(
    x = "kinv",
    y = "train_error"
)
ctest = alt.Chart(df_scores).mark_line(color="orange").encode(
    x = "kinv",
    y = "test_error"
)
ctest+ctrain

Section 5: Candlestick Chart

Instead of using the Scatter() plot, we will create fully interactive chart using plotly. By passing the Close or adjClose price to the y-axis, and we need to specify each of open, high, low, and close by using add_trace(go.Candlestick(...)). With make_subplots(), we also have another smaller chart at the bottom; this is called a “range slider”, and we can drag either side to zoom in/out on a certain chart area. We could hide this range slider by using figure.update_layout(xaxis_rangeslider_visible = False). In the chart below, different colors help distinguish between an up or down day – green for up days and red for down days.

import plotly.graph_objects as go
from plotly.subplots import make_subplots


stock = make_subplots(
            specs=[[{"secondary_y":True}]]
        ).add_trace(go.Candlestick(
                        x=df.date, 
                        open= df["open"], 
                        high=df["high"], 
                        low=df["low"], 
                        close= df["close"],
                        name = 'Price'),
        ).update_layout(title = {"text":"Apple Stock", "x":0.5})
    
stock

Summary

In this project, I started by cleaning and organizing the data set. Then I constructed a few charts to observe and compare the trend of Apple stock performance and displayed the volume distribution over the last few years. Then, I calculated the 15-day Exponential Moving Average using the collected data. Using linear regression with EMA15 and Adjust Close data to predict future stock prices. I observed a DataFrame of train error and test error with a range of k from 1 to 150 by using the K-Nearest Neighbors Regressor. Finally, I created a candlestick chart with open, high, low, and close values.

References

  • What is the source of your dataset(s)?

    • Kaggle Dataset

Created in deepnote.com Created in Deepnote

previous

Depression Prediction

next

Analysis of credit card fraud