Updated the Script for User Friendliness

BTCUSDT_predictions.csv

@@ -0,0 +1,31 @@
+,Model 1,Model 2,Model 3,Model 4
+0,63919.35250000008,64262.7,63908.37834900037,63840.04536619295
+1,63423.25429999992,63152.01,63659.1707101622,63869.0389345954
+2,65011.0950000001,65173.99,65014.57449067909,64552.76661555629
+3,65660.47850000008,65769.95,65870.51656146956,65699.04049608468
+4,65769.97089999999,65858.0,65773.77101660996,65868.47218892295
+5,65565.34639999991,65602.01,65613.15961152391,65846.58070227425
+6,63570.34319999994,63327.59,63641.94846348572,64265.206929840926
+7,61297.37070000002,60805.78,61771.543690837556,62176.306500999766
+8,60944.10550000003,60649.28,61534.934782153476,61374.58113794234
+9,60738.67029999998,60752.71,60688.189725915814,60712.640070678615
+10,61761.47590000002,62086.0,61937.26222859245,61660.19218442663
+11,62065.076300000015,62058.0,61899.53361315821,62009.28258436733
+12,62632.12420000006,62819.91,62527.3250838097,62461.66321806782
+13,62498.7231,62224.0,63600.12523207192,63355.616200261706
+14,62263.82590000009,62160.49,62636.08780019019,62521.448614304965
+15,60854.69439999994,60636.02,61071.78310790617,61360.46045651742
+16,60342.635399999934,60326.39,59908.79966274048,60026.65307039782
+17,62291.43989999998,62540.0,62675.830201943,61950.334281343516
+18,63145.033999999985,63206.22,63081.48889013403,62890.30474420209
+19,62705.33739999991,62870.02,62258.20326144328,62468.333053275805
+20,65821.8547000001,66083.99,65589.0249382245,64837.914767528484
+21,67016.85639999995,67074.14,66671.4704391068,66593.74607362071
+22,67729.3118999999,67620.01,67829.04104327965,67725.99301787547
+23,67412.25690000004,67421.78,67059.39845251033,67327.04957688058
+24,68286.07669999998,68428.0,68435.02049263423,68223.88474096541
+25,68516.87600000002,68378.0,68160.33558887534,68324.97185214274
+26,68882.26350000009,69031.99,68870.69962816169,68802.72221711298
+27,67630.0641,67377.5,67766.39120386967,68104.99534838935
+28,67346.69970000001,67426.0,67081.28148397617,67282.25415273676
+29,66685.0858,66473.0,66525.79468248309,66951.00763572242

README.md

@@ -1,6 +1,6 @@
 # Crypto Price Prediction Bot
 
-Crypto Price Prediction Bot is a Telegram bot that predicts the price of cryptocurrencies using historical data from the Binance exchange. It utilizes machine learning models to analyze the data and generate price predictions for the next 30 days.
+The **Crypto Price Prediction Bot** is a Telegram bot designed to predict the prices of cryptocurrencies using historical data from the Binance exchange. It leverages machine learning models to analyze data and generate price predictions for the next 30 days, along with interactive visualizations.
 
 ## Getting Started
 
@@ -19,10 +19,23 @@ To get started with the Crypto Price Prediction Bot, you need to set up a Telegr
 5. Enter the symbol of the cryptocurrency you want to predict (e.g., BTCUSDT).
 6. The bot will process the data, train the models, and generate price predictions.
 7. The bot will send you a CSV file containing the predictions for the next 30 days.
+8. **Download CSV File**: The bot will send you a CSV file containing the predictions for the next 30 days.
+9. **View Interactive Graphs**: The bot will also generate interactive graphs displaying the predictions from different models.
+
+## Features
+
+- **Data Retrieval**: Fetches historical price data from the Binance exchange.
+- **Data Cleaning**: Cleans and processes the data by calculating RSI, Moving Average, and MACD indicators.
+- **Machine Learning Models**: Trains multiple models including Random Forest, Decision Tree, Linear Regression, and Ridge regression.
+- **Price Prediction**: Predicts prices for the next 30 days based on historical data.
+- **Interactive Visualizations**: Uses Plotly to create interactive plots for a more engaging user experience.
+
+![Screenshot 2024-10-23 165824](https://github.com/user-attachments/assets/6dd17aca-82bc-4ddd-aa80-3e94bc93884e)
+
 
 ## File Description
 
-- `bot.py`: The main Python script that implements the Telegram bot and handles user interactions.
+- `main.py`: The main Python script that implements the Telegram bot and handles user interactions.
 - `ccxt`: A Python library for cryptocurrency trading and market data retrieval from various exchanges.
 - `numpy`: A library for scientific computing with Python.
 - `pandas`: A library for data manipulation and analysis.
@@ -33,6 +46,7 @@ To get started with the Crypto Price Prediction Bot, you need to set up a Telegr
 ## Contributors
 
 - [Sheekovic](https://github.com/Sheekovic)
+- [Aas1kk](https://github.com/aa-sikkkk)
 
 Feel free to contribute to the project by creating issues or submitting pull requests.
 

main.py

@@ -2,6 +2,7 @@
 import ccxt
 import numpy as np
 import pandas as pd
+import plotly.graph_objects as go
 from sklearn.model_selection import train_test_split
 from sklearn.preprocessing import StandardScaler
 from sklearn.impute import SimpleImputer
@@ -15,7 +16,7 @@
 from telebot.types import Message
 
 # Initialize the Telegram bot
-bot = telebot.TeleBot('PUT YOUR BOT TOKEN HERE')
+bot = telebot.TeleBot('YOUR_TOKEN_HERE')
 print('Bot started')
 
 # Function to get the crypto symbol from the user
@@ -26,49 +27,30 @@ def get_crypto_symbol(message: Message):
 # Import Bitcoin Price Data from Binance
 def get_binance_data(crypto_symbol):
     print(f"Getting {crypto_symbol} data from Binance...")
-    # Create a Binance exchange instance
     binance = ccxt.binance()
-
-    # Define the symbol and timeframe
     symbol = crypto_symbol
     timeframe = '1d'
-
-    # Fetch OHLCV (Open-High-Low-Close-Volume) data
     ohlcv = binance.fetch_ohlcv(symbol, timeframe, limit=1000)
 
-    # if data is not empty
     if ohlcv:
-        # Create a DataFrame
         df = pd.DataFrame(ohlcv, columns=['Open Time', 'Open', 'High', 'Low', 'Close', 'Volume'])
-
-        # Convert the timestamp to datetime
         df['Open Time'] = pd.to_datetime(df['Open Time'], unit='ms')
-
-        # Set the index to the date column
         df.set_index('Open Time', inplace=True)
-
-        # Convert the columns to numeric
         df = df.apply(pd.to_numeric)
-
-        # Return the DataFrame
         return df
     else:
         print('No data found')
         return None
 
 # Save the data to a CSV file
 def save_data(df, crypto_symbol):
-    # Save the data to a CSV file
     df.to_csv(f'{crypto_symbol}.csv')
     print(f"Data saved to {crypto_symbol}.csv")
-
     os.remove(f'{crypto_symbol}.csv')
 
 # Load the data from a CSV file
 def load_data(crypto_symbol):
-    # Load the data from a CSV file
     df = pd.read_csv(f'{crypto_symbol}.csv', index_col='Open Time', parse_dates=True)
-    # Return the DataFrame
     return df
 
 # Calculate RSI
@@ -84,117 +66,140 @@ def calculate_rsi(data, window=14):
 
 # Clean the data
 def clean_data(df):
-    # Drop the rows with missing values
     df.dropna(inplace=True)
-    # Drop the duplicate rows
     df.drop_duplicates(inplace=True)
-
-    # Calculate RSI
     df['RSI'] = calculate_rsi(df['Close'], window=14)
-
-    # Calculate Moving Average
     df['MA'] = df['Close'].rolling(window=20).mean()
-
-    # Calculate MACD
     macd = MACD(df['Close'], window_slow=26, window_fast=12, window_sign=9)
     df['MACD'] = macd.macd()
-
-    # Return the cleaned DataFrame
     return df
 
 # Train the model with the data
 def train_model(df):
-    # Create the features and target
     X = df.drop('Close', axis=1)
     y = df['Close']
-    # Split the data into training and testing sets
     X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.2, random_state=42)
 
-    # Define the column transformer
     column_transformer = make_column_transformer(
         (SimpleImputer(), X_train.columns),
         remainder='passthrough'
     )
 
-    # Create a list of models
     models = [
         ('Random Forest', make_pipeline(column_transformer, RandomForestRegressor())),
         ('Decision Tree', make_pipeline(column_transformer, DecisionTreeRegressor())),
         ('Linear Regression', make_pipeline(column_transformer, LinearRegression())),
         ('Ridge', make_pipeline(column_transformer, Ridge()))
     ]
 
-    # Create a dictionary to store the model scores
     model_scores = {}
-
-    # Loop through the models
     for name, model in models:
-        # Fit the model to the data
         model.fit(X_train, y_train)
-        # Evaluate the model and store its score
         score = model.score(X_test, y_test)
         model_scores[name] = score
-
-    # Return the model scores
     return model_scores
 
-
 # Print the model scores
 def print_model_scores(model_scores):
-    # Loop through the model names and scores and print them
     for i, (name, score) in enumerate(model_scores.items()):
-        model_name = list(model_scores.keys())[i]  # Extract the model name from the keys
+        model_name = list(model_scores.keys())[i]
         print(f'{model_name} Model R-squared: {score*100:.2f}%')
 
 # Predict the price for the next 30 days
 def predict_price(df):
-    # Create the features and target
     X = df.drop('Close', axis=1)
     y = df['Close']
 
-    # Impute missing values
     imputer = SimpleImputer()
     X = imputer.fit_transform(X)
 
-    # Scale the data
     scaler = StandardScaler()
     scaler.fit(X)
     X = scaler.transform(X)
 
-    # Create a list of models
     models = [
         ('Random Forest', RandomForestRegressor()),
         ('Decision Tree', DecisionTreeRegressor()),
         ('Linear Regression', LinearRegression()),
         ('Ridge', Ridge())
     ]
-    # Create a list to store the predictions
+
     predictions = []
-    # Loop through the models
     for name, model in models:
-        # Fit the model to the data
         model.fit(X, y)
-        # Predict the price for the next 30 days
         prediction = model.predict(X[-30:])
-        # Append the predictions to the list
         predictions.append(prediction)
-    # Convert the predictions list to a numpy array
+
     predictions = np.array(predictions).T
-    # Return the predictions
     return predictions
 
 # Save the predictions to a symbol name.csv file
 def save_predictions(predictions, crypto_symbol, file_name):
-    # Create a DataFrame
     df = pd.DataFrame(predictions)
-    # Get the model names based on the number of columns
     model_names = [f'Model {i+1}' for i in range(predictions.shape[1])]
-    # Set the columns of the DataFrame
     df.columns = model_names
-    # Save the predictions to a CSV file
     df.to_csv(file_name)
     print(f"Predictions saved to {file_name}")
 
+# Advanced interactive plot with Plotly (Dropdown, Range Slider, Annotations)
+def plot_predictions_advanced(crypto_symbol, file_name):
+    predictions_df = pd.read_csv(file_name)
+
+    # Create a figure
+    fig = go.Figure()
+
+    # Add traces for each model's predictions
+    for column in predictions_df.columns[1:]:
+        fig.add_trace(go.Scatter(
+            x=predictions_df.index,
+            y=predictions_df[column],
+            mode='lines+markers',
+            name=column,
+            visible=True if column == 'Model 1' else False  # Set only Model 1 visible initially
+        ))
+
+    # Dropdown menu for selecting models to display
+    dropdown_buttons = []
+    for column in predictions_df.columns[1:]:
+        dropdown_buttons.append(dict(
+            args=[{'visible': [col == column for col in predictions_df.columns[1:]]}],
+            label=column,
+            method='restyle'
+        ))
+
+    fig.update_layout(
+        updatemenus=[
+            dict(
+                active=0,
+                buttons=dropdown_buttons,
+                x=1.15,  # Place it to the right
+                y=0.5
+            )
+        ]
+    )
+
+    # Range slider
+    fig.update_layout(
+        xaxis=dict(
+            rangeslider=dict(
+                visible=True
+            )
+        )
+    )
+
+    # Add title and labels
+    fig.update_layout(
+        title=f'Predictions for {crypto_symbol} - Model Comparison',
+        xaxis_title='Days',
+        yaxis_title='Predicted Price',
+        template='plotly_white',
+        hovermode='x unified',
+        legend_title_text='Models',
+    )
+
+    # Show the interactive plot in the browser or supported terminal
+    fig.show()
+
 # Handle the '/start' command
 @bot.message_handler(commands=['start'])
 def handle_start(message):
@@ -203,37 +208,31 @@ def handle_start(message):
 # Handle the '/predict' command
 @bot.message_handler(commands=['predict'])
 def handle_predict(message):
-    # Ask the user for the symbol of the cryptocurrency
     bot.reply_to(message, "Enter the symbol of the cryptocurrency (e.g., BTCUSDT):")
-    # Register the next handler to receive the symbol from the user
     bot.register_next_step_handler(message, process_crypto_symbol)
 
-
 # Process the cryptocurrency symbol
 def process_crypto_symbol(message):
     crypto_symbol = message.text
-    # Get the data from Binance
     data = get_binance_data(crypto_symbol)
     if data is not None:
-        # Clean the data
         cleaned_data = clean_data(data)
-        # Train the model
         model_scores = train_model(cleaned_data)
-        # Print the model scores
         print_model_scores(model_scores)
-        # Predict the price for the next 30 days
         predictions = predict_price(cleaned_data)
-        # Save the predictions to a CSV file
+
+        # Save CSV in both Telegram and root folder
         file_name = f'{crypto_symbol}_predictions.csv'
         save_predictions(predictions, crypto_symbol, file_name)
-        # Send the predictions file to the user
+
+        # Send the CSV to Telegram chat
         bot.send_document(message.chat.id, open(file_name, 'rb'))
-        # Delete the CSV file
-        os.remove(file_name)
+
+        # Plot predictions interactively
+        plot_predictions_advanced(crypto_symbol, file_name)
     else:
         bot.reply_to(message, "No data found for the given cryptocurrency symbol.")
 
-
 # Handle all other messages
 @bot.message_handler(func=lambda message: True)
 def handle_message(message):