Great job so far! Let's break down **your code, the error, and how to solve it**, and then I'll also address both questions (1: hourly, 2: daily aggregation and prediction). --- ## **Your Error Explained** Your error: ``` Error in plot.default(...) : formal argument "type" matched by multiple actual arguments ``` **Reason:** You are calling `plot(BikeDemand, type = "l", ...)` where `BikeDemand` is a **data frame** with 7 columns. The base `plot()` function for data frames calls `plot.data.frame()`, which does **not** have a `type` argument. The `type` argument is only valid for `plot.default()` (i.e., for vectors or matrices). **To plot all 6 stations as lines, you should:** 1. Extract the time and station columns. 2. Use `matplot()` for multiple time series. 3. Or loop through columns and use `lines()`. --- ### **A. HOW TO FIX THE PLOT** #### **Option 1: Use `matplot`** ```r # Assume first column is time, next 6 are stations station_data <- as.matrix(BikeDemand[, 2:7]) hours <- BikeDemand[[1]] matplot(hours, station_data, type = "l", lty = 1, col = 1:6, xlab = "Hours", ylab = "Bike Usage/Demand", main = "Types of moving averages", sub = "Oct 3 2016--Oct 23 2016") legend("topright", legend = names(BikeDemand)[2:7], col = 1:6, lty = 1) ``` #### **Option 2: Use ggplot2 (Recommended for R)** ```r library(reshape2) df_long <- melt(BikeDemand, id.vars = names(BikeDemand)[1], variable.name = "Station", value.name = "Demand") library(ggplot2) ggplot(df_long, aes_string(x = names(BikeDemand)[1], y = "Demand", color = "Station")) + geom_line() + labs(title = "Types of moving averages", subtitle = "Oct 3 2016--Oct 23 2016", x = "Hours", y = "Bike Usage/Demand") ``` --- ## **B. Predicting the Next 3 Hours (Question 1)** ### **Simple Approach: Moving Average Prediction** If you want to use your `movavg` function for simple forecasting: ```r # Predict next 3 hours for each station last_n <- 3 # or whatever window you prefer predict_next_hours <- function(x, n = last_n) { rep(mean(tail(x, n)), 3) } predictions <- apply(BikeDemand[, 2:7], 2, predict_next_hours) colnames(predictions) <- names(BikeDemand)[2:7] rownames(predictions) <- paste0("t+", 1:3) predictions ``` - This simply takes the mean of the last `n` hours as the prediction for the next 3 hours. **Discussion:** - This assumes stationarity and no trend/seasonality, which is a *very* simple model. More advanced: use ARIMA (see below). --- ### **C. Aggregating to Daily Data and Predicting Next 3 Days (Question 2)** #### **Aggregate Hourly Data to Daily** ```r BikeDemand$day <- rep(1:21, each = 24)[1:nrow(BikeDemand)] daily_usage <- aggregate(. ~ day, BikeDemand[, -1], FUN = sum) # Now daily_usage has 21 days x 6 stations ``` #### **Predict Next 3 Days** ```r # Again, use simple mean of last 3 days daily_preds <- apply(daily_usage[, -1], 2, function(x) rep(mean(tail(x, 3)), 3)) colnames(daily_preds) <- names(daily_usage)[-1] rownames(daily_preds) <- paste0("Day+", 1:3) daily_preds ``` --- ## **D. (Optional) Advanced Forecast: Use ARIMA** If you want a real time series forecast, use `forecast` package: ```r library(forecast) # For a single station, e.g., station 2 fit <- auto.arima(BikeDemand[[2]]) future <- forecast(fit, h = 3) plot(future) future$mean # predicted next 3 hours # For daily fit_daily <- auto.arima(daily_usage[[2]]) future_daily <- forecast(fit_daily, h = 3) plot(future_daily) future_daily$mean ``` (Repeat for each station.) --- ## **Summary of What You Did Wrong** **Problem:** You called `plot()` with a data frame and a `type` argument, which isn't valid for data frames. **Solution:** - Use `matplot()` or `ggplot2` for timeseries by station. - For forecasting, use your moving average or, better, ARIMA models. --- ## **Full Code Example for Plotting and Simple Prediction** ```r library(readxl) BikeDemand <- read_excel("C:/Users/16092/Desktop/CitiBike_Oct2016_BikeDemand.xlsx", range = "A1:G505") # Plot station_data <- as.matrix(BikeDemand[, 2:7]) hours <- BikeDemand[[1]] matplot(hours, station_data, type = "l", lty = 1, col = 1:6, xlab = "Hours", ylab = "Bike Usage/Demand", main = "Types of moving averages", sub = "Oct 3 2016--Oct 23 2016") legend("topright", legend = names(BikeDemand)[2:7], col = 1:6, lty = 1) # Predict next 3 hours predict_next_hours <- function(x, n = 3) { rep(mean(tail(x, n)), 3) } hourly_pred <- apply(BikeDemand[, 2:7], 2, predict_next_hours) # Aggregate to daily BikeDemand$day <- rep(1:21, each = 24)[1:nrow(BikeDemand)] daily_usage <- aggregate(. ~ day, BikeDemand[, -1], FUN = sum) # Predict next 3 days predict_next_days <- function(x, n = 3) { rep(mean(tail(x, n)), 3) } daily_pred <- apply(daily_usage[, -1], 2, predict_next_days) # Print predictions hourly_pred daily_pred ``` --- **If you want to use ARIMA for better forecasts, let me know and I can provide that code as well.** --- **Summary:** - Fix plotting with `matplot` or `ggplot2`. - For prediction, use moving average or ARIMA for more accuracy. - Aggregate with `aggregate()` for daily. --- **If you share your data structure or want code for specific models, just ask!**