feat: add article

fiqgant · Jul 16, 2024 · f396d4f · f396d4f
1 parent 99ebc81
commit f396d4f
Show file tree

Hide file tree

Showing 5 changed files with 314 additions and 0 deletions.
diff --git a/.../images/blog/bitcoin-price-prediction-using-arima/Plot the time series data.png b/.../images/blog/bitcoin-price-prediction-using-arima/Plot the time series data.png
diff --git a/apps/web/public/images/blog/bitcoin-price-prediction-using-arima/cover.png b/apps/web/public/images/blog/bitcoin-price-prediction-using-arima/cover.png
diff --git a/...c/images/blog/bitcoin-price-prediction-using-arima/first-order differencing.png b/...c/images/blog/bitcoin-price-prediction-using-arima/first-order differencing.png
diff --git a/apps/web/public/images/blog/bitcoin-price-prediction-using-arima/forecast.png b/apps/web/public/images/blog/bitcoin-price-prediction-using-arima/forecast.png
diff --git a/apps/web/src/content/blog/bitcoin-price-prediction-using-arima.mdx b/apps/web/src/content/blog/bitcoin-price-prediction-using-arima.mdx
@@ -0,0 +1,314 @@
+---
+title: Bitcoin Price Prediction Using ARIMA
+date: '2024-07-16T00:00:00Z'
+modifiedTime: '2024-07-16T00:00:00Z'
+summary: Learn how to use ARIMA to predict Bitcoin prices and convert the forecasted values to USD and IDR.
+---
+
+## Introduction
+
+Predicting Bitcoin prices is a challenging yet fascinating task due to the cryptocurrency's volatile nature. In this tutorial, we will guide you through the process of building a time series forecasting model using ARIMA to predict Bitcoin prices and convert the forecasted values to USD and IDR.
+
+## Packages Required
+
+To get started, you'll need the following packages:
+
+- [pandas](https://pandas.pydata.org/) for data manipulation
+- [numpy](https://numpy.org/) for numerical operations
+- [statsmodels](https://www.statsmodels.org/stable/index.html) for statistical modeling
+- [matplotlib](https://matplotlib.org/) for plotting
+- [requests](https://docs.python-requests.org/en/master/) for fetching exchange rate data
+
+You can install these packages in Google Colab using the following commands:
+
+```bash title="Install Packages in Google Colab"
+!pip install pandas numpy statsmodels matplotlib requests
+```
+
+## Step-by-Step Guide
+
+### Step 1: Load the Data
+
+First, we'll load the latest Bitcoin price dataset from Yahoo Finance.
+
+```python title="Load Bitcoin Price Data"
+import pandas as pd
+
+# Load the dataset
+url = 'https://query1.finance.yahoo.com/v7/finance/download/BTC-USD?period1=1451606400&period2=9999999999&interval=1d&events=history'
+data = pd.read_csv(url, parse_dates=['Date'], index_col='Date')
+
+# Display the first few rows
+print(data.head())
+```
+
+```bash title="Output"
+                   Open        High         Low       Close   Adj Close  \
+Date
+2016-01-01  430.721008  436.246002  427.515015  434.334015  434.334015
+2016-01-02  434.622009  436.062012  431.869995  433.437988  433.437988
+2016-01-03  433.578003  433.743011  424.705994  430.010986  430.010986
+2016-01-04  430.061005  434.516998  429.084015  433.091003  433.091003
+2016-01-05  433.069000  434.182007  429.675995  431.959991  431.959991
+
+                Volume
+Date
+2016-01-01  36278900.0
+2016-01-02  30096600.0
+2016-01-03  39633800.0
+2016-01-04  38477500.0
+2016-01-05  34522600.0
+```
+
+### Step 2: Visualize the Data
+
+Visualizing the time series data helps to understand the trends, seasonality, and patterns.
+
+```python title="Visualize Bitcoin Price Data"
+import matplotlib.pyplot as plt
+
+# Plot the time series data
+plt.figure(figsize=(10, 6))
+plt.plot(data['Close'], label='Bitcoin Price')
+plt.title('Daily Bitcoin Prices')
+plt.xlabel('Date')
+plt.ylabel('Price (USD)')
+plt.legend()
+plt.show()
+```
+
+<Image
+  src='images/blog/bitcoin-price-prediction-using-arima/Plot the time series data.png'
+  width={1920}
+  height={1024}
+  alt='https://colab.research.google.com/drive/1O24WTBXypHPeagpLmJyC7DEjCNbKFBjr?usp=sharing'
+/>
+
+<LinkCard
+  href='https://colab.research.google.com/drive/1O24WTBXypHPeagpLmJyC7DEjCNbKFBjr?usp=sharing'
+  hostname='https://colab.research.google.com/'
+  title='Google Colab'
+/>
+
+### Step 3: Differencing the Data
+
+Differencing is used to make the time series data stationary, which is required for ARIMA modeling.
+
+```python title="Perform Differencing on Data"
+# Perform first-order differencing
+data_diff = data['Close'].diff().dropna()
+
+# Plot the differenced data
+plt.figure(figsize=(10, 6))
+plt.plot(data_diff, label='Differenced Bitcoin Price')
+plt.title('Differenced Daily Bitcoin Prices')
+plt.xlabel('Date')
+plt.ylabel('Differenced Price (USD)')
+plt.legend()
+plt.show()
+```
+
+<Image
+  src='images/blog/bitcoin-price-prediction-using-arima/first-order differencing.png'
+  width={1920}
+  height={1024}
+  alt='https://colab.research.google.com/drive/1O24WTBXypHPeagpLmJyC7DEjCNbKFBjr?usp=sharing'
+/>
+
+<LinkCard
+  href='https://colab.research.google.com/drive/1O24WTBXypHPeagpLmJyC7DEjCNbKFBjr?usp=sharing'
+  hostname='https://colab.research.google.com/'
+  title='Google Colab'
+/>
+
+### Step 4: Fit the ARIMA Model
+
+Now, we can fit the ARIMA model to the differenced data. We'll use the ARIMA function from the statsmodels library.
+
+```python title="Fit ARIMA Model"
+from statsmodels.tsa.arima.model import ARIMA
+
+# Fit the ARIMA model
+model = ARIMA(data['Close'], order=(5, 1, 0))  # (p, d, q) parameters
+model_fit = model.fit()
+
+# Print the model summary
+print(model_fit.summary())
+```
+
+```bash title="Output"
+                               SARIMAX Results
+==============================================================================
+Dep. Variable:                  Close   No. Observations:                 3120
+Model:                 ARIMA(5, 1, 0)   Log Likelihood              -25710.140
+Date:                Tue, 16 Jul 2024   AIC                          51432.279
+Time:                        03:16:46   BIC                          51468.551
+Sample:                    01-01-2016   HQIC                         51445.300
+                         - 07-16-2024
+Covariance Type:                  opg
+==============================================================================
+                 coef    std err          z      P>|z|      [0.025      0.975]
+------------------------------------------------------------------------------
+ar.L1         -0.0513      0.010     -5.339      0.000      -0.070      -0.032
+ar.L2          0.0122      0.010      1.198      0.231      -0.008       0.032
+ar.L3          0.0329      0.011      2.974      0.003       0.011       0.055
+ar.L4          0.0363      0.009      3.887      0.000       0.018       0.055
+ar.L5          0.0114      0.009      1.214      0.225      -0.007       0.030
+sigma2      8.492e+05   8408.771    100.987      0.000    8.33e+05    8.66e+05
+===================================================================================
+Ljung-Box (L1) (Q):                   0.03   Jarque-Bera (JB):             16897.89
+Prob(Q):                              0.86   Prob(JB):                         0.00
+Heteroskedasticity (H):              13.15   Skew:                            -0.04
+Prob(H) (two-sided):                  0.00   Kurtosis:                        14.40
+===================================================================================
+
+Warnings:
+[1] Covariance matrix calculated using the outer product of gradients (complex-step).
+```
+
+### Step 5: Forecasting
+
+After fitting the model, we can use it to make forecasts.
+
+```python title="Forecast Bitcoin Prices"
+# Make forecast
+forecast = model_fit.forecast(steps=30)  # Forecasting for the next 30 days
+
+# Plot the forecast
+plt.figure(figsize=(10, 6))
+plt.plot(data['Close'], label='Original')
+plt.plot(forecast, label='Forecast (USD)', color='red')
+plt.title('Bitcoin Price Forecast')
+plt.xlabel('Date')
+plt.ylabel('Price (USD)')
+plt.legend()
+plt.show()
+```
+
+<Image
+  src='images/blog/bitcoin-price-prediction-using-arima/forecast.png'
+  width={1920}
+  height={1024}
+  alt='https://colab.research.google.com/drive/1O24WTBXypHPeagpLmJyC7DEjCNbKFBjr?usp=sharing'
+/>
+
+<LinkCard
+  href='https://colab.research.google.com/drive/1O24WTBXypHPeagpLmJyC7DEjCNbKFBjr?usp=sharing'
+  hostname='https://colab.research.google.com/'
+  title='Google Colab'
+/>
+
+### Step 6: Convert Forecast to IDR
+
+For converting the forecasted Bitcoin prices to IDR, we need the current exchange rate. We'll use the requests library to fetch the latest exchange rate from an API.
+
+```python title="Convert Forecast to IDR"
+import requests
+
+# Function to get USD to IDR exchange rate
+def get_usd_to_idr_exchange_rate():
+    try:
+        response = requests.get('https://api.exchangerate-api.com/v4/latest/USD')
+        data = response.json()
+        return data['rates']['IDR']
+    except Exception as e:
+        print(f"Error getting exchange rate: {e}")
+        return 14500  # Default exchange rate if API fails
+
+# Get the exchange rate
+exchange_rate = get_usd_to_idr_exchange_rate()
+forecast_idr = forecast * exchange_rate
+
+# Print forecasted values in USD and IDR
+print("Forecasted Bitcoin Prices for the next 30 days (USD):")
+print(forecast)
+print("\nForecasted Bitcoin Prices for the next 30 days (IDR):")
+print(forecast_idr)
+```
+
+```bash title="Output"
+Forecasted Bitcoin Prices for the next 30 days (USD):
+2024-07-17    64857.928145
+2024-07-18    65018.188590
+2024-07-19    65167.803824
+2024-07-20    65261.151784
+2024-07-21    65288.089636
+2024-07-22    65298.841536
+2024-07-23    65308.944573
+2024-07-24    65314.532450
+2024-07-25    65316.760778
+2024-07-26    65317.743490
+2024-07-27    65318.393172
+2024-07-28    65318.762795
+2024-07-29    65318.928459
+2024-07-30    65319.006839
+2024-07-31    65319.051756
+2024-08-01    65319.076660
+2024-08-02    65319.088722
+2024-08-03    65319.094613
+2024-08-04    65319.097798
+2024-08-05    65319.099518
+2024-08-06    65319.100384
+2024-08-07    65319.100816
+2024-08-08    65319.101043
+2024-08-09    65319.101164
+2024-08-10    65319.101226
+2024-08-11    65319.101257
+2024-08-12    65319.101273
+2024-08-13    65319.101282
+2024-08-14    65319.101286
+2024-08-15    65319.101289
+Freq: D, Name: predicted_mean, dtype: float64
+
+Forecasted Bitcoin Prices for the next 30 days (IDR):
+2024-07-17    1.050802e+09
+2024-07-18    1.053398e+09
+2024-07-19    1.055822e+09
+2024-07-20    1.057334e+09
+2024-07-21    1.057771e+09
+2024-07-22    1.057945e+09
+2024-07-23    1.058109e+09
+2024-07-24    1.058199e+09
+2024-07-25    1.058235e+09
+2024-07-26    1.058251e+09
+2024-07-27    1.058262e+09
+2024-07-28    1.058268e+09
+2024-07-29    1.058270e+09
+2024-07-30    1.058272e+09
+2024-07-31    1.058272e+09
+2024-08-01    1.058273e+09
+2024-08-02    1.058273e+09
+2024-08-03    1.058273e+09
+2024-08-04    1.058273e+09
+2024-08-05    1.058273e+09
+2024-08-06    1.058273e+09
+2024-08-07    1.058273e+09
+2024-08-08    1.058273e+09
+2024-08-09    1.058273e+09
+2024-08-10    1.058273e+09
+2024-08-11    1.058273e+09
+2024-08-12    1.058273e+09
+2024-08-13    1.058273e+09
+2024-08-14    1.058273e+09
+2024-08-15    1.058273e+09
+Freq: D, Name: predicted_mean, dtype: float64
+```
+
+## Conclusion
+
+This tutorial provided a basic introduction to using ARIMA for time series forecasting of Bitcoin prices. Additionally, we demonstrated how to convert the forecasted values to USD and IDR using the current exchange rate.
+
+## Useful Links
+
+- [ARIMA Model](https://www.statsmodels.org/stable/generated/statsmodels.tsa.arima.model.ARIMA.html)
+- [Time Series Analysis](https://www.statsmodels.org/stable/tsa.html)
+- [Pandas Documentation](https://pandas.pydata.org/pandas-docs/stable/)
+- [Matplotlib Documentation](https://matplotlib.org/)
+- [Requests Documentation](https://docs.python-requests.org/en/master/)
+
+By following these steps, you should be able to create a time series forecasting model using ARIMA and apply it to your own datasets, including Bitcoin prices.
+
+## Links
+
+- [Google Colab Notebook](https://colab.research.google.com/drive/your_google_colab_link_here)
+- [GitHub Repository](https://github.com/your_github_repository_link_here)