Top 10 Cryptos Price Predictions and Forecasts¶

TABLE OF CONTENTS¶

DATA PREPARATION
1. Cryptocurrency Market Data
FIND CORRELATIONS
1. Correlations Plot
BUILDING MODEL
RESULT

In [1]:

import pandas as pd
import numpy as np
import matplotlib.pyplot as plt
import os
import json
import time
import requests
import seaborn as sns
import plotly.graph_objects as go
import plotly.express as px
%matplotlib inline

DATA PREPARATION ¶

Cryptocurrency Market Data ¶

In [172]:

ohlcv_data = pd.read_csv('data/cryptos-ohlcv-daily.csv')
ohlcv_data.head()

Out[172]:

	open	high	low	close	volume	weightedAverage	date	id	symbol
0	27210.36	27350.00	26605.05	26817.93	39705.76020	27014.145	2023-06-01	215	BTC
1	26817.93	27300.00	26505.00	27242.59	36765.30191	27030.260	2023-06-02	215	BTC
2	27242.59	27333.29	26914.93	27069.22	16326.78504	27155.905	2023-06-03	215	BTC
3	27069.22	27455.02	26951.00	27115.20	18888.19016	27092.210	2023-06-04	215	BTC
4	27115.14	27129.33	25388.00	25728.20	69948.65543	26421.670	2023-06-05	215	BTC

Missing Data ¶

In [ ]:

ohlcv_data = ohlcv_data.dropna() 
print(ohlcv_data.isnull().sum())

open               0
high               0
low                0
close              0
volume             0
weightedAverage    0
date               0
id                 0
symbol             0
dtype: int64
open               0
high               0
low                0
close              0
volume             0
weightedAverage    0
date               0
id                 0
symbol             0
dtype: int64

Explore Daily Price Trends ¶

In [174]:

# Convert 'date' column to datetime for better plotting
symbols = ohlcv_data['symbol'].unique()

# Loop through each symbol and create a candlestick plot
for symbol in symbols:
    # Filter data for the current symbol
    symbol_data = ohlcv_data[ohlcv_data['symbol'] == symbol]
    
    # Create the candlestick plot
    fig = go.Figure(data=[go.Candlestick(
        x=symbol_data['date'],
        open=symbol_data['open'],
        high=symbol_data['high'],
        low=symbol_data['low'],
        close=symbol_data['close']
    )])
    
    # Update the layout for better visualization
    fig.update_layout(
        title=f'{symbol} Daily Candlestick Chart',
        xaxis_title='Date',
        yaxis_title='Price (USD)',
        xaxis_rangeslider_visible=False,  # Optional: Hide range slider
        template="plotly_white",
        xaxis=dict(
            tickformat='%b %Y',  # Show month and year
            tickvals=pd.date_range(start=symbol_data['date'].min(), end=symbol_data['date'].max(), freq='MS'),  # Monthly ticks
            tickangle=-45
        ),
    )
    
    # Show the plot for the current symbol
    fig.show()

Explore Top 10 Cryptos Popularity ¶

In [175]:

global_min = ohlcv_data['weightedAverage'].min()
global_max = ohlcv_data['weightedAverage'].max()

ohlcv_data['weighted_score'] = ohlcv_data['weightedAverage'].apply(
    lambda x: 1 + (x - global_min) / (global_max - global_min) * 9
)

# Create a new DataFrame with unique symbols and their normalized weighted scores
unique_crypto_weights = ohlcv_data[['symbol', 'weighted_score']].drop_duplicates(subset='symbol')

display(unique_crypto_weights)

fig = px.bar(
    unique_crypto_weights,
    x='symbol',
    y='weighted_score',
    text='weighted_score',
    color='weighted_score',
    color_continuous_scale=px.colors.sequential.YlOrBr,
    title='Popular Cryptocurrency Weight Distribution',
    labels={'symbol': 'Cryptocurrency', 'weighted_score': 'Weight'}
)

fig.update_traces(
    texttemplate='%{text:.2f}',  # Show text inside the bars
    textposition='inside'
)
fig.update_layout(
    width=800, 
    height=600,
    xaxis=dict(title='Cryptocurrency', tickangle=-45),  
    yaxis=dict(title='Weight'),
    coloraxis_colorbar=dict(title="Weight"), 
    title=dict(font=dict(size=18), x=0.5), 
    plot_bgcolor='rgba(240,240,240,0.5)' 
)

# Show the plot
fig.show()

	symbol	weighted_score
0	BTC	3.466140
553	ETH	1.170499
1106	BNB	1.027916
1659	SOL	1.001880
2212	XRP	1.000041
2765	TRX	1.000002
3318	AVAX	1.001281
3871	DOGE	1.000001
4424	DOT	1.000475
4977	LTC	1.008421

FIND CORRELATIONS ¶

In [176]:

numerical_columns = ['open', 'high', 'low', 'close', 'volume']
numerical_data = ohlcv_data[numerical_columns]
correlation_matrix = numerical_data.corr()
print(correlation_matrix)

            open      high       low     close    volume
open    1.000000  0.999775  0.999777  0.999597 -0.104230
high    0.999775  1.000000  0.999699  0.999862 -0.104174
low     0.999777  0.999699  1.000000  0.999797 -0.104254
close   0.999597  0.999862  0.999797  1.000000 -0.104201
volume -0.104230 -0.104174 -0.104254 -0.104201  1.000000

Correlations Plot ¶

In [177]:

# Plot the heatmap
plt.figure(figsize=(10, 8))
sns.heatmap(correlation_matrix, annot=True, cmap='coolwarm', fmt=".2f", linewidths=0.5)
plt.title('Correlation Matrix for Cryptocurrency Data', fontsize=16)
plt.show()

BUILDING MODEL ¶

Feature Selection ¶

In [178]:

model_data = ohlcv_data

model_data['ma_5'] = model_data.groupby('symbol')['close'].transform(lambda x: x.rolling(window=5).mean())
model_data['ma_20'] = model_data.groupby('symbol')['close'].transform(lambda x: x.rolling(window=20).mean())

# Add lagged features
for lag in range(1, 4):  # Lagged values for the past 3 days
    model_data[f'lag_{lag}'] = model_data.groupby('symbol')['close'].shift(lag)

# Add daily return as a feature
model_data['daily_return'] = model_data.groupby('symbol')['close'].pct_change()

model_data = model_data.dropna()
# Columns used to train the model (e.g., open, high, low, volume, etc.).
features = [
    'open', 'high', 'low', 'volume', 'ma_5', 'ma_20', 
    'lag_1', 'lag_2', 'lag_3'
]

print(model_data)

          open      high       low     close        volume  weightedAverage  \
19    28308.00  30800.00  28257.99  29993.89  1.142210e+05        29150.945   
20    29993.89  30500.00  29525.61  29884.92  6.331282e+04        29939.405   
21    29884.91  31431.94  29800.00  30688.50  8.158243e+04        30286.705   
22    30688.51  30800.00  30250.00  30527.43  3.201427e+04        30607.970   
23    30527.44  31046.01  30277.49  30462.66  3.155572e+04        30495.050   
...        ...       ...       ...       ...           ...              ...   
5525    102.75    122.25     99.35    119.57  1.807382e+06          111.160   
5526    119.57    139.42    114.92    132.93  4.045921e+06          126.250   
5527    132.75    135.16    122.70    130.89  1.956176e+06          131.820   
5528    130.90    135.62    121.97    133.69  1.384338e+06          132.295   
5529    133.69    146.79    124.94    135.78  2.098452e+06          134.735   

            date   id symbol  weighted_score       ma_5       ma_20     lag_1  \
19    2023-06-21  215    BTC        3.661211  27600.640  26635.0170  28307.99   
20    2023-06-22  215    BTC        3.733190  28274.226  26788.3665  29993.89   
21    2023-06-23  215    BTC        3.764895  29143.932  26960.6620  29884.92   
22    2023-06-24  215    BTC        3.794224  29880.546  27133.5725  30688.50   
23    2023-06-25  215    BTC        3.783915  30311.480  27300.9455  30527.43   
...          ...  ...    ...             ...        ...         ...       ...   
5525  2024-12-01  887    LTC        1.010143    103.970     92.1560    102.74   
5526  2024-12-02  887    LTC        1.011520    111.110     94.9680    119.57   
5527  2024-12-03  887    LTC        1.012029    118.180     97.7430    132.93   
5528  2024-12-04  887    LTC        1.012072    123.964    100.3085    130.89   
5529  2024-12-05  887    LTC        1.012295    130.572    102.9125    133.69   

         lag_2     lag_3  daily_return  
19    26844.36  26339.97      0.059556  
20    28307.99  26844.36     -0.003633  
21    29993.89  28307.99      0.026889  
22    29884.92  29993.89     -0.005249  
23    30688.50  29884.92     -0.002122  
...        ...       ...           ...  
5525    104.77     95.54      0.163812  
5526    102.74    104.77      0.111734  
5527    119.57    102.74     -0.015346  
5528    132.93    119.57      0.021392  
5529    130.89    132.93      0.015633  

[5340 rows x 16 columns]

Handling Missing Values ¶

In [179]:

# Check for missing values in features
print("Missing values in features:")
timeframes = [5, 30, 90, 180, 365]

# Fill NaN in target columns with the last known price
for t in timeframes:
    model_data[f'target_{t}d'] = model_data.groupby('symbol')['close'].transform(
        lambda x: x.shift(-t).ewm(span=t, adjust=False).mean()
    )

# Replace NaN values with the last known price
last_known_prices = model_data.groupby('symbol')['close'].last()

# Step 2: Forward fill missing values within each symbol group
for t in timeframes:
    model_data[f'target_{t}d'] = model_data.groupby('symbol')[f'target_{t}d'].fillna(method='ffill')

# Step 3: Fill remaining NaN values with the last known price for each symbol
for t in timeframes:
    model_data[f'target_{t}d'] = model_data.apply(
        lambda row: last_known_prices[row['symbol']] if pd.isna(row[f'target_{t}d']) else row[f'target_{t}d'], axis=1
    )

Missing values in features:

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:7: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:7: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:7: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:7: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:7: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: FutureWarning:

SeriesGroupBy.fillna is deprecated and will be removed in a future version. Use obj.ffill() or obj.bfill() for forward or backward filling instead. If you want to fill with a single value, use Series.fillna instead

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: FutureWarning:

Series.fillna with 'method' is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: FutureWarning:

SeriesGroupBy.fillna is deprecated and will be removed in a future version. Use obj.ffill() or obj.bfill() for forward or backward filling instead. If you want to fill with a single value, use Series.fillna instead

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: FutureWarning:

Series.fillna with 'method' is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: FutureWarning:

SeriesGroupBy.fillna is deprecated and will be removed in a future version. Use obj.ffill() or obj.bfill() for forward or backward filling instead. If you want to fill with a single value, use Series.fillna instead

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: FutureWarning:

Series.fillna with 'method' is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: FutureWarning:

SeriesGroupBy.fillna is deprecated and will be removed in a future version. Use obj.ffill() or obj.bfill() for forward or backward filling instead. If you want to fill with a single value, use Series.fillna instead

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: FutureWarning:

Series.fillna with 'method' is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: FutureWarning:

SeriesGroupBy.fillna is deprecated and will be removed in a future version. Use obj.ffill() or obj.bfill() for forward or backward filling instead. If you want to fill with a single value, use Series.fillna instead

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: FutureWarning:

Series.fillna with 'method' is deprecated and will raise in a future version. Use obj.ffill() or obj.bfill() instead.

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:16: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/2843346195.py:20: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

Train-Test Split ¶

80-20 split: 80% of the data (292 days) is used for training, and 20% (73 days) for testing.

The 80-20 split is a standard practice in machine learning and data analysis.

We used time-based split instead of random selection and it is the best practice for time-series data because:

It preserves the sequential nature of the data.
This simulates a real-world scenario where you train the model on historical data and then evaluate or use it to predict future data.

In [181]:

# Train-test split
from datetime import datetime, timedelta

model_data['date'] = pd.to_datetime(model_data['date'])

start_date = datetime(2023, 6, 1)
end_date = datetime(2024, 12, 5)
total_days = (end_date - start_date).days + 1 # Total number of days including the end date
training_days = int(total_days * 0.8) # Calculate the number of training days (80% of total days)
cutoff_date = pd.to_datetime((start_date + timedelta(days=training_days - 1)))# Calculate the cutoff date

train_data = model_data[model_data['date'] < cutoff_date] # Contains rows before the cutoff date.
test_data = model_data[model_data['date'] >= cutoff_date] # Contains rows from the cutoff date onward.

targets = [f'target_{t}d' for t in timeframes] # The prediction targets (target_5d, target_30d, etc.) defined for different timeframes (e.g., 5 days, 1 month, etc.).

X_train = train_data[features]
X_test = test_data[features]
Y_train = train_data[targets]
Y_test = test_data[targets]


print(X_train.shape)
print(X_test.shape)
print(Y_train.shape)
print(Y_test.shape)

(4220, 9)
(1120, 9)
(4220, 5)
(1120, 5)

/var/folders/96/zkx_7hk579v922k6dk_fc26c0000gn/T/ipykernel_26365/1468742617.py:4: SettingWithCopyWarning:


A value is trying to be set on a copy of a slice from a DataFrame.
Try using .loc[row_indexer,col_indexer] = value instead

See the caveats in the documentation: https://pandas.pydata.org/pandas-docs/stable/user_guide/indexing.html#returning-a-view-versus-a-copy

Model Selection ¶

In [182]:

# X_train, X_test: Feature sets.
# Y_train, Y_test: Target values for each timeframe.
#Train Model
from sklearn.ensemble import RandomForestRegressor # A machine learning model used for regression tasks.

# Train models for each target
models = {}
predictions = {}

for target in Y_train.columns: # Loops through each target column (target_5d, target_30d, etc.).
    # Train the model
    model = RandomForestRegressor(n_estimators=100, random_state=42) # Trains a separate Random Forest model for each target timeframe.
    model.fit(X_train, Y_train[target])
    
    # Save the model
    models[target] = model # Each trained model is stored in the models dictionary for future use.
    
    # Make predictions
    predictions[target] = model.predict(X_test) # Predictions for the test set (X_test) are stored in the predictions dictionary.

Model Evaluation ¶

In [183]:

from sklearn.metrics import mean_squared_error

# Mean Squared Error (MSE): A common metric for evaluating regression models.
# Measures the average squared difference between the actual (y_test) and predicted values (predictions[target]).
# Calculates and prints the MSE for each prediction target (target_5d, target_30d, etc.).
# A lower MSE indicates better model performance.
for target in predictions.keys():
    mse = mean_squared_error(Y_test[target], predictions[target])
    print(f"{target} - MSE: {mse:.2f}")

target_5d - MSE: 20720289.35
target_30d - MSE: 28110491.94
target_90d - MSE: 20997786.79
target_180d - MSE: 3297805.00
target_365d - MSE: 4080.68

RESULTS ¶

In [186]:

# Summarize predictions by symbol

summary = model_data.groupby('symbol').last().reset_index()
# Select relevant columns for summary
summary = summary[['symbol', 'close', 'target_5d', 'target_30d',
                   'target_90d', 'target_180d', 'target_365d']].sort_values("close", ascending= False)

# Rename columns for clarity
summary.rename(columns={
    'name': 'Name',
    'symbol': 'Symbol',
    'category': 'Crypto Type',
    'description': 'Description',
    'close': 'Current Price',
    'target_5d': '5D Prediction',
    'target_30d': '1M Prediction',
    'target_90d': '3M Prediction',
    'target_180d': '6M Prediction',
    'target_365d': '1Y Prediction',
}, inplace=True)

# Format prices as currency
summary['Current Price'] = summary['Current Price'].apply(lambda x: f"${x:,.2f}")
summary['5D Prediction'] = summary['5D Prediction'].apply(lambda x: f"${x:,.2f}")
summary['1M Prediction'] = summary['1M Prediction'].apply(lambda x: f"${x:,.2f}")
summary['3M Prediction'] = summary['3M Prediction'].apply(lambda x: f"${x:,.2f}")
summary['6M Prediction'] = summary['6M Prediction'].apply(lambda x: f"${x:,.2f}")
summary['1Y Prediction'] = summary['1Y Prediction'].apply(lambda x: f"${x:,.2f}")

summary.reset_index(drop=True, inplace=True)

pd.set_option('display.max_colwidth', 100) 
pd.set_option('display.max_rows', None) 
pd.set_option('display.max_columns', None) 
display(summary)

	Symbol	Current Price	5D Prediction	1M Prediction	3M Prediction	6M Prediction	1Y Prediction
0	BTC	$96,945.63	$96,939.33	$90,473.03	$77,968.14	$70,912.84	$67,982.73
1	ETH	$3,785.20	$3,725.93	$3,337.72	$2,985.61	$2,926.36	$3,122.91
2	BNB	$717.10	$704.77	$645.32	$605.68	$576.16	$583.39
3	SOL	$235.84	$233.65	$223.79	$191.07	$171.69	$155.72
4	LTC	$135.78	$128.49	$99.34	$82.41	$77.36	$74.38
5	AVAX	$50.30	$49.92	$40.31	$32.95	$31.50	$28.39
6	DOT	$10.37	$9.98	$7.51	$5.88	$5.73	$5.43
7	XRP	$2.24	$2.29	$1.50	$0.97	$0.78	$0.64
8	DOGE	$0.43	$0.42	$0.36	$0.25	$0.19	$0.16
9	TRX	$0.32	$0.31	$0.22	$0.18	$0.16	$0.14

In [189]:

import matplotlib.pyplot as plt

# Define the list of coins
coins = summary['Symbol'].unique()  # Adjusted to match "Name" column in the table

# Define the timeframes and labels
timeframes = ['5D Prediction', '1M Prediction', '3M Prediction', '6M Prediction', '1Y Prediction']
labels = ['5D', '1M', '3M', '6M', '1Y']

# Create a 2x5 grid for the plots
fig, axes = plt.subplots(2, 5, figsize=(20, 10))  # 2 rows, 5 columns
axes = axes.flatten()  # Flatten axes array for easy iteration

# Plot data for each coin
for i, coin in enumerate(coins):
    # Extract data for the current coin
    coin_data = summary[summary['Symbol'] == coin].iloc[0]
    predictions = coin_data[timeframes].values

    # Convert predictions to float for plotting (if they are strings with "$" formatting)
    predictions = [float(pred.replace('$', '').replace(',', '')) for pred in predictions]

    # Plot the predictions
    axes[i].plot(labels, predictions, marker='o', linestyle='-', color='blue')
    axes[i].set_title(f'{coin} Predicted Prices', fontsize=10)
    axes[i].set_ylabel('Predicted Price', fontsize=8)
    axes[i].set_xlabel('Timeframe', fontsize=8)
    axes[i].grid(True)

    # Customize y-axis for better readability
    axes[i].tick_params(axis='y', labelsize=8)
    axes[i].tick_params(axis='x', labelsize=8)

# Adjust layout and remove empty subplots
for ax in axes[len(coins):]:  # Hide any unused subplots
    ax.axis('off')

# Adjust layout for better spacing
plt.tight_layout()

# Display the plot
plt.show()

In [192]:

#Top 10 Cryptocurrency Price Predictions Summary Plot

import matplotlib.pyplot as plt

# Define the timeframes and labels
timeframes = ['5D Prediction', '1M Prediction', '3M Prediction', '6M Prediction', '1Y Prediction']
labels = ['5D', '1M', '3M', '6M', '1Y']

# Create the plot
# Create the plot
plt.figure(figsize=(12, 8))

# Plot data for each coin
for coin in summary['Symbol'].unique():
    coin_data = summary[summary['Symbol'] == coin].iloc[0]
    predictions = coin_data[timeframes].values
    predictions = [float(pred.replace('$', '').replace(',', '')) for pred in predictions]
    plt.plot(labels, predictions, marker='o', label=coin)

# Add title and labels
plt.title("Top 10 Cryptocurrency Price Predictions (Log Scale)", fontsize=14)
plt.xlabel("Timeframe", fontsize=12)
plt.ylabel("Predicted Price (USD)", fontsize=12)

# Set Y-axis to logarithmic scale
plt.yscale('log')

# Add a legend
plt.legend(title="Cryptocurrency", fontsize=10, bbox_to_anchor=(1.05, 1), loc='upper left')

# Customize grid
plt.grid(axis='y', linestyle='--', alpha=0.7)

# Adjust layout for better spacing
plt.tight_layout()

# Display the plot
plt.show()

Top 10 Cryptos Price Predictions and Forecasts¶

TABLE OF CONTENTS¶

DATA PREPARATION¶

Cryptocurrency Market Data¶

Missing Data¶

Explore Daily Price Trends¶

Explore Top 10 Cryptos Popularity¶

FIND CORRELATIONS¶

Correlations Plot¶

BUILDING MODEL¶

Feature Selection¶

Handling Missing Values¶

Train-Test Split¶

Model Selection¶

Model Evaluation¶

RESULTS¶