tech knights Tshilisanani Dzivhani Prayer Nkuna

solution.r

@@ -0,0 +1,127 @@
+# Load required libraries
+library(tidyverse)
+library(lubridate)
+library(plotly)
+
+print(paste("---> R Script Start", Sys.time()))
+
+# Set up parameters
+start_train <- as.Date("2017-01-01")
+end_train <- as.Date("2023-11-30")
+start_test <- as.Date("2024-01-01")
+end_test <- as.Date("2024-06-30")
+
+n_buys <- 10
+lookback_period <- 20  # Number of days to look back for momentum calculation
+
+# Load data
+df_sectors <- read.csv('data/data0.csv')
+df_data <- read.csv('data/data1.csv')
+df_data$date <- as.Date(df_data$date)
+
+# Function to calculate momentum
+calculate_momentum <- function(df) {
+  df %>%
+    group_by(security) %>%
+    arrange(date) %>%
+    mutate(momentum = (price / lag(price, lookback_period) - 1) * 100) %>%
+    ungroup()
+}
+
+# Prepare data
+df_momentum <- df_data %>%
+  select(date, security, price) %>%
+  calculate_momentum()
+
+# Modified generate_buy_signals function
+generate_buy_signals <- function(df, n_buys) {
+  df %>%
+    arrange(desc(momentum)) %>%
+    slice_head(n = n_buys) %>%
+    mutate(signal = 1) %>%
+    select(security, signal)
+}
+
+# Generate signals for all test dates
+df_signals <- df_momentum %>%
+  filter(date >= start_test, date <= end_test) %>%
+  group_by(date) %>%
+  group_modify(~ generate_buy_signals(.x, n_buys)) %>%
+  ungroup()
+
+# Create buy matrix
+df_buys <- df_signals %>%
+  pivot_wider(names_from = security, values_from = signal, values_fill = 0)
+
+# Plot heatmap of buy signals
+fig_pixel <- plot_ly(z = as.matrix(df_buys[,-1]), x = colnames(df_buys)[-1], y = df_buys$date, type = "heatmap")
+fig_pixel
+
+# Load returns data
+df_returns <- read.csv('data/returns.csv')
+df_returns$date <- as.Date(df_returns$date)
+df_returns <- df_returns %>% 
+  filter(date >= start_test) %>%
+  pivot_wider(names_from = security, values_from = return1)
+
+# Updated calculate_portfolio_performance function
+calculate_portfolio_performance <- function(df_buys, df_returns) {
+  # Ensure df_buys and df_returns have the same columns, excluding 'date'
+  common_cols <- setdiff(intersect(names(df_buys), names(df_returns)), "date")
+  print("Common columns (excluding 'date'):")
+  print(common_cols)
+
+  df_buys <- df_buys[, c("date", common_cols)]
+  df_returns <- df_returns[, c("date", common_cols)]
+
+  # Aggregate data by date to ensure uniqueness
+  df_buys <- df_buys %>%
+    group_by(date) %>%
+    summarise(across(everything(), sum)) %>%
+    ungroup()
+  print("Aggregated df_buys:")
+  print(head(df_buys))
+
+  df_returns <- df_returns %>%
+    group_by(date) %>%
+    summarise(across(everything(), mean)) %>%
+    ungroup()
+  print("Aggregated df_returns:")
+  print(head(df_returns))
+
+  # Align dates
+  df_combined <- inner_join(df_buys, df_returns, by = "date", suffix = c("_buys", "_returns"))
+  print("Combined data frame:")
+  print(head(df_combined))
+
+  # Calculate daily portfolio return
+  df_combined$portfolio_return <- rowSums(df_combined[, paste0(common_cols, "_buys")] * 
+                                            df_combined[, paste0(common_cols, "_returns")], na.rm = TRUE) / n_buys
+  print("Data frame with portfolio return:")
+  print(head(df_combined))
+
+  # Calculate cumulative return
+  df_combined$cumulative_return <- cumprod(1 + df_combined$portfolio_return)
+  print("Data frame with cumulative return:")
+  print(head(df_combined))
+
+  return(df_combined)
+}
+
+# Use the updated function
+df_performance <- calculate_portfolio_performance(df_buys, df_returns)
+
+# Plot performance
+fig_performance <- plot_ly(df_performance, x = ~date, y = ~cumulative_return, type = 'scatter', mode = 'lines')
+fig_performance
+
+# Print final performance
+final_return <- tail(df_performance$cumulative_return, 1)
+print(paste("---> Total return for the momentum strategy between", 
+            min(df_performance$date), "and", max(df_performance$date), 
+            "is", round((final_return - 1) * 100, 2), "%"))
+
+print(paste("---> R Script End", Sys.time()))
+print(paste("---> Total time taken", difftime(Sys.time(), as.POSIXct(start_train), units = "secs")))
+
+