0

SOLVED

I'm working on replication of the code that is used at the Reproducible Finance with R. A link to the webinar is here: https://www.rstudio.com/resources/webinars/reproducible-finance-with-r/

The data for the exercise was downloaded from Yahoo Finance. The .csv file is here: http://www.reproduciblefinance.com/data/data-download/

Following the instructions provided at the webinar, it happened to me that the code doesn't work as it is in the lesson:

portfolio_returns_tidyquant_rebalanced_monthly %>%  
        mutate(
                dplyr_port_returns = portfolio_returns_dplyr_byhand$returns,
                xts_port_returns = coredata(portfolio_returns_xts_rebalanced_monthly)
                )%>%
        head()

The system doesn't provide any output, nor it shows if there is a mistake in the code.

I then decided to eliminate each new variable I want to create to see if something happens. It turned out that if one variable is not included in the mutate() command the system produces a partial output I need. Below are a code and an output.

portfolio_returns_tidyquant_rebalanced_monthly %>%  
        mutate(
                dplyr_port_returns = portfolio_returns_dplyr_byhand$returns,
              #  xts_port_returns = coredata(portfolio_returns_xts_rebalanced_monthly)
                )%>%
        head(2)
Date returns dplyr_port_returns
2013-01-31 0.0308487341 0.0308487341
2013-02-28 -0.0008697461 -0.0008697461

In addition, some information about variables:

class(portfolio_returns_xts_rebalanced_monthly)
[1] "xts" "zoo"

class(portfolio_returns_dplyr_byhand)
[1] "tbl_df"     "tbl"        "data.frame"

The portfolio_returns_xts_rebalanced_monthly was created using the following code:

symbols <- c("SPY", "EFA", "IJS", "EEM", "AGG")

prices <-
        getSymbols(
                symbols,
                src = 'yahoo',
                from = "2012-12-31",
                to = "2017-12-31",
                auto.assign = T,
                warnings = F
        ) %>%
        map(~Ad(get(.))) %>% 
        reduce(merge) %>% 
        `colnames<-`(symbols)
w <- c(
        0.25,
        0.25,
        0.20,
        0.20,
        0.10
)

prices_monthly <-
        to.monthly(
                prices,
                indexAt = "lastof", 
                OHLC = FALSE
        )

assets_return_xts <- na.omit(
        Return.calculate(
                prices_monthly,
                method = "log"
        )
)

portfolio_returns_xts_rebalanced_monthly <- 
        Return.portfolio(
                assets_return_xts,
                weights = w,
                rebalance_on = 'months'
        ) %>% 
        `colnames<-`("returns")

I'm pretty sure this is somehow connected to a mutate() function and classes of variables, but I couldn't find any information on the matter. Your support is highly appreciated.

UPDATE.

Changing the class of one object from xts to data.frame, and adjusting a code a bit solved the issue.

An updated code:

portfolio_returns_xts_rebalanced_monthly_df <-
  data.frame(
    date=index(portfolio_returns_xts_rebalanced_monthly), 
    coredata(portfolio_returns_xts_rebalanced_monthly)
    )

portfolio_returns_tidyquant_rebalanced_monthly %>%  
        mutate(
                dplyr_port_returns = portfolio_returns_dplyr_byhand$returns,
                xts_port_returns = portfolio_returns_xts_rebalanced_monthly_df$returns
                )%>%
        head()
dkolkin
  • 81
  • 10
  • 1
    Are you sure that the output of your `coredata` command is a vector? That's what `mutate` expects. – Claudio May 20 '21 at 12:01
  • @RonakShah, I edited the message accordingly. Thank you, I missed this point. – dkolkin May 20 '21 at 12:21
  • @Claudio, I checked, and it is not a vector. – dkolkin May 20 '21 at 12:23
  • 1
    That's probably why your code fails. The data you point to is a dataframe, but `portfolio_returns_xts_rebalanced_monthly` is an `xts` object. How did you create that object? You need to give us all the data we need to reproduce the problem. – Claudio May 20 '21 at 12:35
  • @Claudio, gotcha. Made edits above. – dkolkin May 20 '21 at 13:43
  • 1
    make the xts object a data.frame first with `df <- data.frame(date=index(portfolio_returns_xts_rebalanced_monthly), coredata(portfolio_returns_xts_rebalanced_monthly))` – MGP May 20 '21 at 13:52
  • Yes, this approach worked. Thank you very much. – dkolkin May 20 '21 at 15:54

0 Answers0