Data Exploration

Quick look at the data

Show the code

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd
import numpy as np

Data Understanding

Show the code

# Apply the default theme
sns.set_theme(style="whitegrid", palette="pastel")

# Load a local file
stocks = pd.read_csv("stock.csv")
print(stocks)

           STLA         GM          TM      NSANY      MBGYY      BMWYY  \
0     11.472710  37.606449  128.369995  20.040001  15.427056  25.132998   
1     11.945811  38.524120  130.130005  20.250000  15.544848  25.270374   
2     12.854664  39.711693  132.160004  20.230000  15.751438  25.443907   
3     13.551867  39.594738  133.860001  20.389999  15.934468  25.740351   
4     13.433591  39.783680  134.770004  20.440001  16.003330  25.913887   
...         ...        ...         ...        ...        ...        ...   
1464  18.480000  28.549999  171.610001   7.780000  15.140000  31.129999   
1465  18.040001  27.219999  172.809998   7.800000  14.710000  31.059999   
1466  18.000000  27.360001  171.399994   7.590000  14.780000  31.219999   
1467  18.680000  28.200001  175.179993   7.690000  14.600000  30.940001   
1468  18.940001  28.000000  185.779999   7.930000  14.790000  31.530001   

         POAHY        TSLA       Dates  
0     6.742718   21.368668  2018-01-02  
1     6.887550   21.150000  2018-01-03  
2     7.072612   20.974667  2018-01-04  
3     7.145028   21.105333  2018-01-05  
4     7.161119   22.427334  2018-01-08  
...        ...         ...         ...  
1464  4.440000  205.759995  2023-10-26  
1465  4.390000  207.300003  2023-10-27  
1466  4.400000  197.360001  2023-10-30  
1467  4.410000  200.839996  2023-10-31  
1468  4.460000  205.660004  2023-11-01  

[1469 rows x 9 columns]

Show the code

stocks.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 1469 entries, 0 to 1468
Data columns (total 9 columns):
 #   Column  Non-Null Count  Dtype  
---  ------  --------------  -----  
 0   STLA    1469 non-null   float64
 1   GM      1469 non-null   float64
 2   TM      1469 non-null   float64
 3   NSANY   1469 non-null   float64
 4   MBGYY   1469 non-null   float64
 5   BMWYY   1469 non-null   float64
 6   POAHY   1469 non-null   float64
 7   TSLA    1469 non-null   float64
 8   Dates   1469 non-null   object 
dtypes: float64(8), object(1)
memory usage: 103.4+ KB

Show the code

stocks['Dates'] = pd.to_datetime(stocks['Dates'])  # Convert the 'Dates' column to datetime format

# Filter the dataframe for only 2023 dates
stocks_2023 = stocks[stocks['Dates'].dt.year == 2023]

print(stocks_2023)

           STLA         GM          TM  NSANY      MBGYY      BMWYY     POAHY  \
1259  13.374442  33.556999  138.279999   6.33  15.643178  27.898378  5.281156   
1260  13.805285  34.420227  137.190002   6.32  16.162148  28.553400  5.405082   
1261  13.906119  34.727821  135.520004   6.31  16.180681  28.793264  5.433681   
1262  14.162791  35.630741  138.970001   6.49  16.421631  29.374483  5.548074   
1263  14.318627  35.640663  139.940002   6.56  16.588442  29.439062  5.633869   
...         ...        ...         ...    ...        ...        ...       ...   
1464  18.480000  28.549999  171.610001   7.78  15.140000  31.129999  4.440000   
1465  18.040001  27.219999  172.809998   7.80  14.710000  31.059999  4.390000   
1466  18.000000  27.360001  171.399994   7.59  14.780000  31.219999  4.400000   
1467  18.680000  28.200001  175.179993   7.69  14.600000  30.940001  4.410000   
1468  18.940001  28.000000  185.779999   7.93  14.790000  31.530001  4.460000   

            TSLA      Dates  
1259  108.099998 2023-01-03  
1260  113.639999 2023-01-04  
1261  110.339996 2023-01-05  
1262  113.059998 2023-01-06  
1263  119.769997 2023-01-09  
...          ...        ...  
1464  205.759995 2023-10-26  
1465  207.300003 2023-10-27  
1466  197.360001 2023-10-30  
1467  200.839996 2023-10-31  
1468  205.660004 2023-11-01  

[210 rows x 9 columns]

Descriptive Statistics

The number of rows and columns

Show the code

# get the number of rows and columns

len(stocks_2023.index), len(stocks_2023.columns)

(210, 9)

NaN values in all columns

Show the code

# Counting NaN values in all columns
nan_count = stocks_2023.isna().sum()

print(nan_count)

STLA     0
GM       0
TM       0
NSANY    0
MBGYY    0
BMWYY    0
POAHY    0
TSLA     0
Dates    0
dtype: int64

Basic Statistics

Show the code

print("----------------------")
print("BASIC STATISTICS: ")
print("----------------------")
stocks_2023.describe()

----------------------
BASIC STATISTICS: 
----------------------

	STLA	GM	TM	NSANY	MBGYY	BMWYY	POAHY	TSLA	Dates
count	210.000000	210.000000	210.000000	210.000000	210.000000	210.000000	210.000000	210.000000	210
mean	17.076711	35.148255	154.424142	7.846190	18.125585	34.915196	5.415873	213.400285	2023-06-03 13:42:51.428571392
min	13.374442	27.219999	133.479996	6.310000	14.600000	27.898378	4.390000	108.099998	2023-01-03 00:00:00
25%	15.901123	32.907166	139.217495	7.372500	17.360720	32.543494	5.219193	183.252499	2023-03-20 06:00:00
50%	16.820590	34.420546	146.985001	7.740000	17.959999	34.445070	5.447980	207.575005	2023-06-03 12:00:00
75%	18.632500	37.777060	167.952503	8.335000	19.015000	37.187501	5.671500	255.487499	2023-08-17 18:00:00
max	20.549999	42.834282	195.039993	9.640000	20.670000	41.189999	6.000000	293.339996	2023-11-01 00:00:00
std	1.692184	3.327676	16.635200	0.698586	1.194182	3.125648	0.356021	44.268249	NaN

Data visualization

Show the code


# Create a visualization
sns.set_theme(style="whitegrid", palette="pastel")

# Create a histogram for TSLA stock prices
sns.lineplot(data=stocks_2023, x='Dates', y='TSLA')
plt.title('TSLA Stock Price Fluctuation for 2023')
plt.xlabel('Dates')
plt.ylabel('Stock Price')
plt.xticks(rotation=45)
plt.show()

In 2023, TSLA’s stock price experienced a sharp rise until mid-year, followed by a decline towards the end of the year.

Show the code

sns.lineplot(data=stocks_2023, x='Dates', y='TSLA', label='TSLA')
sns.lineplot(data=stocks_2023, x='Dates', y='GM', label='GM')
plt.title('TSLA vs GM Stock Price Comparison for 2023')
plt.xlabel('Dates')
plt.ylabel('Stock Price')
plt.legend()
plt.xticks(rotation=45)
plt.show()

Throughout 2023, TSLA’s stock price exhibited significant volatility, peaking around mid-year before declining, while GM’s stock remained relatively stable, hovering at a much lower price range. The comparison shows a stark contrast in the stock price trajectories of the two companies during this period.

Correlation Analysis

Show the code

# INSERT CODE TO RE-CREATE THE FOLLOWING PLOTS 
# SEE IF THERE IS ANY CORRELATION BETWEEN THE CONTINOUS VARIABLES

# INSERT CODE

# LINEAR MODEL PLOT (lmplot --> Plot data and regression model fits across a FacetGrid)
sns.lmplot(data=stocks_2023, x='TSLA', y='STLA')

/Users/isfarbaset/anaconda3/lib/python3.11/site-packages/seaborn/axisgrid.py:118: UserWarning: The figure layout has changed to tight
  self._figure.tight_layout(*args, **kwargs)

The scatter plot displays a positive correlation between “TSLA” and “STLA” stock values. As the “TSLA” value increases on the x-axis, the “STLA” value typically rises on the y-axis, with some variability.

Show the code

# # Convert the dataframe from wide to long format
# # df_melted = stocks_2023.melt(id_vars=['Dates'], value_vars=['STLA', 'GM', 'TM', 'NSANY', 'MBGYY', 'BMWYY', 'POAHY', 'TSLA'], 
# #                     var_name='Brand', value_name='Stock Price')

# print(df_melted)

Show the code

# INSERT CODE TO RE-CREATE THE FOLLOWING PLOTS 

# INSERT CODE

sns.set_theme(style="white")
corr = stocks_2023.corr(numeric_only=True)  #Compute the correlation matrix

# # Generate a mask for the upper triangle
mask = np.triu(np.ones_like(corr, dtype=bool)) 
f, ax = plt.subplots(figsize=(7, 5)) #initialize figure

cmap = sns.diverging_palette(230, 20, as_cmap=True) #custom diverging colormap

# # Draw the heatmap with the mask and correct aspect ratio
sns.heatmap(corr, mask=mask, cmap=cmap, vmax=.3, center=0,
            square=True, linewidths=.5, cbar_kws={"shrink": .5})
plt.show()

Based on the correlation heatmap, it can be said that the stock prices of the different care brands are mostly negatively correlated.

Hypothesis Generation

Based on the different visualizations above, an apt hypothesis to test out could be “TSLA’s significant price surge in 2023 suggests a dominant market influence, potentially driving similar trends in associated automotive stocks”

Data Grouping and Segmentation

Show the code

from sklearn.preprocessing import StandardScaler
from sklearn.cluster import KMeans
import seaborn as sns
import matplotlib.pyplot as plt

stocks_2023 = stocks_2023.drop('Dates', axis=1)

# Standardizing the data (important for K-Means)
scaler = StandardScaler()
scaled_df = scaler.fit_transform(stocks_2023.dropna())

# Using K-Means clustering
kmeans = KMeans(n_clusters=3)  # assuming 3 clusters
stocks_2023['Cluster'] = kmeans.fit_predict(scaled_df)

# Visualize using TSLA and another stock, e.g., 'GM'
sns.lmplot(x='TSLA', y='GM', data=stocks_2023, hue='Cluster', fit_reg=False)
plt.show()

/Users/isfarbaset/anaconda3/lib/python3.11/site-packages/sklearn/cluster/_kmeans.py:1412: FutureWarning: The default value of `n_init` will change from 10 to 'auto' in 1.4. Set the value of `n_init` explicitly to suppress the warning
  super()._check_params_vs_input(X, default_n_init=10)
/Users/isfarbaset/anaconda3/lib/python3.11/site-packages/seaborn/axisgrid.py:118: UserWarning: The figure layout has changed to tight
  self._figure.tight_layout(*args, **kwargs)

The visualization displays three distinct clusters of data points based on TSLA and GM stock values. Cluster 0 primarily represents data points where both TSLA and GM have lower stock values, Cluster 1 captures instances with mid-range values, and Cluster 2 comprises of higher TSLA stock values with varying GM values. The spread of the clusters suggests a potential relationship or pattern between the stock values of TSLA and GM over the period of 2023

Identifying Outliers

Show the code

# Source: https://www.datacamp.com/tutorial/naive-bayes-scikit-learn

def show_boxplot(df):
    plt.rcParams['figure.figsize'] = [14,6]
    sns.boxplot(data = df, orient="v")
    plt.title("Outliers Distribution", fontsize = 16)
    plt.ylabel("Range", fontweight = 'bold')
    plt.xlabel("Attributes", fontweight = 'bold')
   
show_boxplot(stocks_2023)

Here we can see the distribution of outliers for various automotive stock attributes. Most stocks, like GM, TM, NSANY, MBGYY, BMWYY, and POAHY, have a narrow range of outlier values, typically below 100. Notably, TM displays a compact outlier range between 100 and 150, while two instances of TSLA stock demonstrate a broader range of outliers, with one reaching up to 300 and another clustering around 150 to 200.

Report and discussion of methods and findings

Topic	Description	Methods
Correlation Matrix	The matrix depicts correlations between various stock tickers. Blue indicates positive correlation, red indicates negative correlation, and white means no correlation. For instance, “STLA” and “GM” have a strong positive correlation while “TSLA” has low to no correlation with most of the other stocks.	This was done using seaborn’s correlation analysis tool
TSLA Stock Price Fluctuation for 2023	The chart shows Tesla’s stock price performance throughout 2023. Tesla’s stock price has seen a significant rise from January to around March/April and then a fluctuation with a general downtrend.	Matplotlib and Seaborn was used
TSLA vs GM Stock Price Comparison for 2023	This line graph compares the stock prices of Tesla (TSLA) and General Motors (GM) for 2023. While TSLA experienced significant fluctuations in its stock price, GM’s stock price remained relatively stable and lower in comparison.	Matplotlib and seaborn was used
Stock Correlation Heatmap	The heatmap visually represents the correlation between various stock tickers. The colors range from blue (positive correlation) to red (negative correlation) with white indicating no correlation. The vertical and horizontal labels denote the stock tickers being compared.	Seaborn was used
Outliers Distribution	The box plot demonstrates the distribution of outliers for different stock tickers. It helps in identifying stocks with higher price volatility and potential risks. TSLA and TM seem to have a more extensive range compared to others like STLA and GM.	Box Plot using Seaborn and Matplotlib
Stock Price Cluster Analysis (TSLA vs GM)	This scatter plot showcases a cluster analysis between TSLA and GM stock prices. It suggests possible correlation patterns or trading strategies by observing how the stocks cluster together in distinct groups. The colors differentiate the clusters.	Scatter Plot using Matplotlib, Seaborn and SKlearn

Tools and Software

Python with libraries like Pandas, Matplotlib, Seaborn, Numpy and SKlearn