Portfolio Efficient Frontier.R

# tickers : stock tickers based on yahoo-finance 
# start_date, end_date: data length
# num_port: the number of portfolios you want to create, normally 5000 

portfolio_assets = function(tickers,start_date,end_date,num_port){ 
  ifelse(!require('pacman'), install.packages('pacman'), library('pacman')) 
  pacman::p_load("tidyquant","plotly","timetk","tidyr",'zeallot','knitr') 
  '%=%' = zeallot::`%<-%` ; '%+%' = paste0 
  options(scipen = 999, warn = -1) 
  
  log_ret_xts = tq_get( 
    tickers, 
    from = start_date, 
    to = end_date, 
    get = 'stock.prices' ) %>% 
    group_by(symbol) %>% 
    tq_transmute(select = adjusted, mutate_fun = periodReturn, period = 'daily', col_rename = 'ret', type = 'log') %>% 
    pivot_wider(names_from = symbol, values_from = ret) %>% 
    na.omit() %>% 
    tk_xts() 
  
  mean_return = colMeans(log_ret_xts) %>% 
    round(.,digits = 5) 
  cov_mat = round(cov(log_ret_xts)*252,digits=4) #annualizing, multiply 252days 
  
  #create 5000 random portfoilo 
  all_wts = matrix(nrow = num_port, ncol = ncol(log_ret_xts)) 
  c(portfolio_returns,portfolio_risk,sharpe_ratio) %=% rep(list(vector("numeric",length = num_port)),3) # random weights and portfolio return, risk and sharpe ratio, hypothesis : risk-free rate is 0 
  
  for (i in 1:num_port) { 
    
    all_wts[i,] = runif(ncol(log_ret_xts)) %>% 
    {./sum(.)} 
    
    portfolio_returns[i] = (1+all_wts[i,] %*% mean_return)^252 -1 
    portfolio_risk[i] = sqrt(t(all_wts[i,]) %*% (cov_mat %*% all_wts[i,])) 
    sharpe_ratio[i] = portfolio_returns[i]/portfolio_risk[i] 
    
    } 
  
  
  all_wts = all_wts %>% tk_tbl() %>% `colnames<-`(tickers %+% "_weights") 
  port_folio_values = tibble(return = portfolio_returns, risk = portfolio_risk, sharpe_ratio = sharpe_ratio) %>% 
    cbind(all_wts,.) #Global Minimal Variance Portfolio 
  
  risk_min = port_folio_values[which.min(port_folio_values$risk),-c(1:ncol(log_ret_xts))] 
  risk_max = port_folio_values[which.max(port_folio_values$risk),-c(1:ncol(log_ret_xts))] 
  tang = port_folio_values[which.max(port_folio_values$sharpe_ratio),-c(1:ncol(log_ret_xts))] 
  wmin = port_folio_values[which.min(port_folio_values$risk),1:ncol(log_ret_xts)] 
  wmax = port_folio_values[which.max(port_folio_values$risk),1:ncol(log_ret_xts)] 
  wtan = port_folio_values[which.max(port_folio_values$sharpe_ratio),1:ncol(log_ret_xts)] 
  
  
  ann = list( 
    x = risk_min$risk, 
    y = risk_min$return, 
    text = "Minimum Variance Portfolio", 
    xref = "x", yref = "y", 
    showarrow = TRUE, arrowhead = 0) 
  
  ann2 = list( 
    x = risk_max$risk, 
    y = risk_max$return, 
    text = "Maximum Variance Portfolio", 
    xref = "x", yref = "y", 
    showarrow = TRUE, 
    arrowhead = 0) 
  
  
  ann3 = list( 
    x = tang$risk,
    y = tang$return, 
    text = "Tangency Portfolio", 
    xref = "x", yref = "y", 
    showarrow = TRUE, arrowhead = 0) 
  
  title1 = list( 
    text = "<b> Portfolio Optimization and Efficient Frontier </b>", 
    xref = "paper", 
    yref = "paper", 
    xanchor = "center", 
    yanchor = "bottom", 
    align = "center", 
    x = 0.5, 
    y = 1, 
    showarrow = FALSE ) 
  
  title2 = list( 
    text = "<b> Portfolio Weights </b>", 
    xref = "paper", 
    yref = "paper", 
    xanchor = "center", 
    yanchor = "bottom", 
    align = "center", 
    x = 0.5, 
    y = 1, 
    showarrow = FALSE ) 
  
  #### plot #### 
  fig1 = port_folio_values %>% 
    plot_ly(
      x=.$risk, 
      y=.$return, 
      type = 'scatter', 
      mode = 'markers', 
      color = .$sharpe_ratio, 
      showlegend=FALSE) %>% 
    add_markers(
      x = risk_min$risk, 
      y = risk_min$return, 
      markers = list(color = "#A52929")) %>% 
    add_markers(
      x = risk_max$risk, 
      y = risk_max$return, 
      markers = list(color = "#A52929")) %>% 
    add_markers(
      x = tang$risk, 
      y = tang$return, 
      markers = list(color = "#A52929")) %>% 
    layout(xaxis = list(title = "\u03C3",zeroline=F), 
           yaxis = list(title = "\u00B5",zeroline=F), 
           annotations = list(title1,ann,ann2,ann3)) %>% 
    colorbar(title = "Sharpe Ratio") 
  
  
  fig2 = wmin %>% 
    round(x=.,digits=5) %>% 
    {plot_ly( 
      x = colnames(.), 
      y = as.numeric(.[1:ncol(log_ret_xts)]), 
      type = 'bar', 
      name = "Minimum Weights", 
      marker = list(color = '#3246AB')) %>% 
        add_trace( 
          y = as.numeric(wmax), 
          name = "Maximum Weights", 
          marker = list(color = '#E82712')) %>% 
        add_trace( y = as.numeric(wtan), 
                   name = "Tangency Weights", 
                   marker = list(color = '#12B7C3')) %>% 
        layout( 
          annotations = title2, 
          yaxis = list(title = "Weights"), 
          barmode = "group" ) } 
  
  
  subplot(fig1, fig2, nrows = 2, titleY= TRUE, titleX = TRUE, margin = 0.1, heights = c(0.5,0.5)) %>% 
    print() 
  
  cat("============ Summary ============","\n", 
      "** Tangency Portfolio", "\n", 
      "- Return : " %+% (tang$return %>% round(.,digits = 4)),"\n", 
      "- Risk : " %+% (tang$risk %>% round(.,digits = 4)), "\n", 
      "- Sharpe Ratio : " %+% round(tang$sharpe_ratio,digits = 4),"\n") 
  t(wtan) %>% round(x=.,digits = 3) %>% `colnames<-`("Weight") %>% kable() %>% print() 
  cat("\n") 
  
  cat("============ Summary ============","\n", 
      "** Minimum Variance Portfolio", "\n", 
      "- Return : " %+% round(risk_min$return,digits = 4),"\n", 
      "- Risk : " %+% round(risk_min$risk,digits = 4), "\n", 
      "- Sharpe Ratio : " %+% (risk_min$sharpe_ratio %>% round(.,digits = 4)),"\n") 
  t(wmin) %>% round(x=.,digits = 3) %>% `colnames<-`("Weight") %>% kable() %>% print() 
  
  }