Spaces:
Sleeping
Sleeping
Update app.py
Browse files
app.py
CHANGED
|
@@ -5,58 +5,75 @@ import pandas as pd
|
|
| 5 |
import matplotlib.pyplot as plt
|
| 6 |
import scipy.optimize as sco
|
| 7 |
|
|
|
|
| 8 |
def get_stock_data(tickers, start, end):
|
| 9 |
data = yf.download(tickers, start=start, end=end)
|
| 10 |
-
|
| 11 |
-
|
| 12 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
| 13 |
return None
|
| 14 |
|
|
|
|
|
|
|
|
|
|
| 15 |
def calculate_returns(data):
|
| 16 |
log_returns = np.log(data / data.shift(1))
|
| 17 |
return log_returns.mean() * 252, log_returns.cov() * 252
|
| 18 |
|
|
|
|
| 19 |
def optimize_portfolio(returns, cov_matrix):
|
| 20 |
num_assets = len(returns)
|
| 21 |
-
|
| 22 |
def sharpe_ratio(weights):
|
| 23 |
portfolio_return = np.dot(weights, returns)
|
| 24 |
portfolio_volatility = np.sqrt(np.dot(weights.T, np.dot(cov_matrix, weights)))
|
| 25 |
-
return -portfolio_return / portfolio_volatility
|
| 26 |
|
| 27 |
constraints = ({'type': 'eq', 'fun': lambda x: np.sum(x) - 1})
|
| 28 |
bounds = tuple((0, 1) for _ in range(num_assets))
|
| 29 |
init_guess = num_assets * [1. / num_assets]
|
| 30 |
-
|
| 31 |
result = sco.minimize(sharpe_ratio, init_guess, method='SLSQP', bounds=bounds, constraints=constraints)
|
| 32 |
-
return result.x
|
| 33 |
|
|
|
|
| 34 |
st.title("Analisis Portofolio Saham Optimal (Model Markowitz)")
|
| 35 |
|
|
|
|
| 36 |
tickers_list = st.text_input("Masukkan ticker saham (contoh: BBCA.JK, TLKM.JK, BBRI.JK)", "BBCA.JK, TLKM.JK, BBRI.JK").split(", ")
|
| 37 |
start_date = st.date_input("Pilih tanggal mulai", pd.to_datetime("2020-01-01"))
|
| 38 |
end_date = st.date_input("Pilih tanggal akhir", pd.to_datetime("2020-12-31"))
|
| 39 |
|
| 40 |
if st.button("Analisis Portofolio"):
|
| 41 |
try:
|
|
|
|
| 42 |
stock_data = get_stock_data(tickers_list, start_date, end_date)
|
| 43 |
-
|
| 44 |
-
if stock_data is None
|
| 45 |
-
st.error("Data tidak ditemukan atau tidak lengkap. Periksa ticker atau tanggal yang dipilih.")
|
| 46 |
-
else:
|
| 47 |
mean_returns, cov_matrix = calculate_returns(stock_data)
|
|
|
|
|
|
|
| 48 |
optimal_weights = optimize_portfolio(mean_returns, cov_matrix)
|
| 49 |
-
|
| 50 |
-
|
| 51 |
-
|
| 52 |
-
|
| 53 |
-
|
| 54 |
-
|
| 55 |
-
|
| 56 |
-
|
| 57 |
-
|
| 58 |
-
|
| 59 |
-
|
| 60 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
| 61 |
except Exception as e:
|
| 62 |
-
st.error(f"Terjadi kesalahan: {e}")
|
|
|
|
| 5 |
import matplotlib.pyplot as plt
|
| 6 |
import scipy.optimize as sco
|
| 7 |
|
| 8 |
+
# Fungsi untuk mengunduh data saham
|
| 9 |
def get_stock_data(tickers, start, end):
|
| 10 |
data = yf.download(tickers, start=start, end=end)
|
| 11 |
+
|
| 12 |
+
if data.empty:
|
| 13 |
+
st.error("Data saham tidak ditemukan. Periksa ticker atau rentang tanggal.")
|
| 14 |
+
return None
|
| 15 |
+
|
| 16 |
+
if 'Adj Close' not in data.columns:
|
| 17 |
+
st.error("Kolom 'Adj Close' tidak ditemukan dalam data yang diunduh.")
|
| 18 |
return None
|
| 19 |
|
| 20 |
+
return data['Adj Close']
|
| 21 |
+
|
| 22 |
+
# Fungsi untuk menghitung return tahunan dan matriks kovarians
|
| 23 |
def calculate_returns(data):
|
| 24 |
log_returns = np.log(data / data.shift(1))
|
| 25 |
return log_returns.mean() * 252, log_returns.cov() * 252
|
| 26 |
|
| 27 |
+
# Fungsi untuk menghitung portofolio optimal
|
| 28 |
def optimize_portfolio(returns, cov_matrix):
|
| 29 |
num_assets = len(returns)
|
| 30 |
+
|
| 31 |
def sharpe_ratio(weights):
|
| 32 |
portfolio_return = np.dot(weights, returns)
|
| 33 |
portfolio_volatility = np.sqrt(np.dot(weights.T, np.dot(cov_matrix, weights)))
|
| 34 |
+
return -portfolio_return / portfolio_volatility # Negatif untuk minimisasi
|
| 35 |
|
| 36 |
constraints = ({'type': 'eq', 'fun': lambda x: np.sum(x) - 1})
|
| 37 |
bounds = tuple((0, 1) for _ in range(num_assets))
|
| 38 |
init_guess = num_assets * [1. / num_assets]
|
| 39 |
+
|
| 40 |
result = sco.minimize(sharpe_ratio, init_guess, method='SLSQP', bounds=bounds, constraints=constraints)
|
| 41 |
+
return result.x if result.success else None
|
| 42 |
|
| 43 |
+
# Streamlit UI
|
| 44 |
st.title("Analisis Portofolio Saham Optimal (Model Markowitz)")
|
| 45 |
|
| 46 |
+
# Input Saham & Tanggal
|
| 47 |
tickers_list = st.text_input("Masukkan ticker saham (contoh: BBCA.JK, TLKM.JK, BBRI.JK)", "BBCA.JK, TLKM.JK, BBRI.JK").split(", ")
|
| 48 |
start_date = st.date_input("Pilih tanggal mulai", pd.to_datetime("2020-01-01"))
|
| 49 |
end_date = st.date_input("Pilih tanggal akhir", pd.to_datetime("2020-12-31"))
|
| 50 |
|
| 51 |
if st.button("Analisis Portofolio"):
|
| 52 |
try:
|
| 53 |
+
# Ambil data saham
|
| 54 |
stock_data = get_stock_data(tickers_list, start_date, end_date)
|
| 55 |
+
|
| 56 |
+
if stock_data is not None:
|
|
|
|
|
|
|
| 57 |
mean_returns, cov_matrix = calculate_returns(stock_data)
|
| 58 |
+
|
| 59 |
+
# Optimasi portofolio
|
| 60 |
optimal_weights = optimize_portfolio(mean_returns, cov_matrix)
|
| 61 |
+
|
| 62 |
+
if optimal_weights is not None:
|
| 63 |
+
st.subheader("Bobot Portofolio Optimal:")
|
| 64 |
+
for i, stock in enumerate(tickers_list):
|
| 65 |
+
st.write(f"{stock}: {optimal_weights[i]:.2%}")
|
| 66 |
+
|
| 67 |
+
# Plot Efficient Frontier
|
| 68 |
+
st.subheader("Efficient Frontier")
|
| 69 |
+
fig, ax = plt.subplots()
|
| 70 |
+
ax.scatter(np.sqrt(np.diag(cov_matrix)), mean_returns, c=mean_returns / np.sqrt(np.diag(cov_matrix)), marker='o')
|
| 71 |
+
ax.set_xlabel("Risiko (Standar Deviasi)")
|
| 72 |
+
ax.set_ylabel("Return Tahunan")
|
| 73 |
+
ax.set_title("Efficient Frontier")
|
| 74 |
+
st.pyplot(fig)
|
| 75 |
+
else:
|
| 76 |
+
st.error("Optimasi portofolio gagal. Coba dengan saham yang berbeda.")
|
| 77 |
+
|
| 78 |
except Exception as e:
|
| 79 |
+
st.error(f"Terjadi kesalahan: {e}")
|