[Eventstudies-commits] r319 - / pkg/data pkg/man

noreply at r-forge.r-project.org noreply at r-forge.r-project.org
Mon May 12 12:50:15 CEST 2014


Author: chiraganand
Date: 2014-05-12 12:50:15 +0200 (Mon, 12 May 2014)
New Revision: 319

Added:
   pkg/data/OtherReturns.rda
   pkg/man/OtherReturns.Rd
Removed:
   pkg/data/AMMData.rda
   pkg/data/EESData.rda
   pkg/data/INR.rda
   pkg/data/MMData.rda
   pkg/data/NiftyIndex.rda
   pkg/man/AMMData.Rd
   pkg/man/EESData.Rd
   pkg/man/INR.Rd
   pkg/man/MMData.Rd
   pkg/man/NiftyIndex.Rd
Modified:
   pkg/man/ees.Rd
   pkg/man/eesPlot.Rd
   pkg/man/eventstudy.Rd
   pkg/man/excessReturn.Rd
   pkg/man/lmAMM.Rd
   pkg/man/makeX.Rd
   pkg/man/manyfirmssubperiod.lmAMM.Rd
   pkg/man/marketResidual.Rd
   pkg/man/phys2eventtime.Rd
   pkg/man/remap.cumprod.Rd
   pkg/man/remap.cumsum.Rd
   pkg/man/remap.event.reindex.Rd
   pkg/man/subperiod.lmAMM.Rd
   todo.org
Log:
Removed unneeded data sets, clubbed all the other returns into one rda, added it's manual.

Deleted: pkg/data/AMMData.rda
===================================================================
(Binary files differ)

Deleted: pkg/data/EESData.rda
===================================================================
(Binary files differ)

Deleted: pkg/data/INR.rda
===================================================================
(Binary files differ)

Deleted: pkg/data/MMData.rda
===================================================================
(Binary files differ)

Deleted: pkg/data/NiftyIndex.rda
===================================================================
(Binary files differ)

Added: pkg/data/OtherReturns.rda
===================================================================
(Binary files differ)


Property changes on: pkg/data/OtherReturns.rda
___________________________________________________________________
Added: svn:mime-type
   + application/x-xz

Deleted: pkg/man/AMMData.Rd
===================================================================
--- pkg/man/AMMData.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/AMMData.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -1,21 +0,0 @@
-\name{AMMData}
-\alias{AMMData}
-\docType{data}
-
-\title{Data set containing firm returns, market returns, currency
-  returns, and call money rate used for AMM estimation}
-
-\description{This data set consists of daily time series for firm
-  returns (Infosys and TCS), market returns (Nifty returns), currency
-  returns (INR/USD), and call money rate. It is used to demonstrate
-  augmented market model estimation.
-
-  The data series is a daily time-series zoo object. The sample range for
-  the data is from 2012-02-01 to 2014-01-31. All series are in per cent.
-}
-
-\usage{data(AMMData)}
-
-\author{Vikram Bahure}
-
-\keyword{AMMData}

Deleted: pkg/man/EESData.Rd
===================================================================
--- pkg/man/EESData.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/EESData.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -1,15 +0,0 @@
-\name{EESData}
-\alias{EESData}
-\docType{data}
-
-\title{Returns data used for extreme events analysis}
-
-\description{This data set is used to demonstrate extreme events study
-  functionality of the package. It contains daily returns data (in per
-  cent) of S&P 500 and the NIFTY Index.}
-
-\usage{data(EESData)}
-
-\author{Chirag Anand}
-
-\keyword{datasets}

Deleted: pkg/man/INR.Rd
===================================================================
--- pkg/man/INR.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/INR.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -1,18 +0,0 @@
-\name{INR}
-\alias{INR}
-\docType{data}
-
-\title{Exchange rate data of Indian Rupee to US Dollar}
-
-\description{
-  A sample of INR/USD rates from 1990 to 2011.
-}
-\usage{data(INR)}
-
-\format{\pkg{zoo}}
-
-\examples{
-data(INR)
-}
-
-\keyword{datasets}

Deleted: pkg/man/MMData.Rd
===================================================================
--- pkg/man/MMData.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/MMData.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -1,21 +0,0 @@
-\name{MMData}
-\alias{MMData}
-
-
-\title{Sample data used for market model examples}
-
-\description{This data is only used for market model examples.}
-
-\usage{data(MMData)}
-
-\format{\pkg{zoo}}
-
-\author{Vikram Bahure}
-
-\examples{
-library(zoo)
-data(MMData)
-str(MMData)
-}
-
-\keyword{MMData}

Deleted: pkg/man/NiftyIndex.Rd
===================================================================
--- pkg/man/NiftyIndex.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/NiftyIndex.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -1,14 +0,0 @@
-\name{NiftyIndex}
-\alias{NiftyIndex}
-\docType{data}
-
-\title{NSE Nifty index from 2004 to 2012}
-
-\description{Time series of Nifty index return (in per cent) from 1990
-  to 2012.}
-
-\usage{data(NiftyIndex)}
-
-\author{Vikram Bahure}
-
-\keyword{NiftyIndex}

Added: pkg/man/OtherReturns.Rd
===================================================================
--- pkg/man/OtherReturns.Rd	                        (rev 0)
+++ pkg/man/OtherReturns.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -0,0 +1,20 @@
+\name{OtherReturns}
+\alias{OtherReturns}
+\docType{data}
+
+\title{Data set containing daily returns of Nifty index, USD INR, call momey
+  rate, and S&P 500 index.}
+
+\description{This data set consists of daily time series of market
+  returns (Nifty index and S&P 500 index), currency returns (USD/INR),
+  and call money rate.
+
+  The data series is a daily time-series zoo object. All series are in
+  per cent.
+}
+
+\usage{data(OtherReturns)}
+
+\author{Chirag Anand}
+
+\keyword{OtherReturns}

Modified: pkg/man/ees.Rd
===================================================================
--- pkg/man/ees.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/ees.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -85,7 +85,8 @@
 \author{Vikram Bahure, Vimal Balasubramaniam}
 
 \examples{
-data(EESData)
-r <- ees(EESData$sp500, prob.value = 5)
+data(OtherReturns)
+
+r <- ees(OtherReturns$SP500, prob.value = 5)
 str(r, max.level = 2)
 }

Modified: pkg/man/eesPlot.Rd
===================================================================
--- pkg/man/eesPlot.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/eesPlot.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -69,10 +69,11 @@
 \author{Vikram Bahure, Vimal Balasubramaniam}
 
 \examples{
-data(EESData)
-eesPlot(z = EESData,
-        response.series.name = "nifty",
-        event.series.name = "sp500",
+data("OtherReturns")
+
+eesPlot(z = OtherReturns,
+        response.series.name = OtherReturns$NiftyIndex,
+        event.series.name = OtherReturns$SP500,
         titlestring = "S&P500",
         ylab = "(Cum.) change in NIFTY")
 }

Modified: pkg/man/eventstudy.Rd
===================================================================
--- pkg/man/eventstudy.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/eventstudy.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -218,12 +218,21 @@
 plot(es)
 
 ## Event study using Augment Market Model
-data("AMMData")
+data("OtherReturns")
+
 events <- data.frame(outcome.unit = c("Infosys", "TCS"),
                      event.when = c("2012-04-01", "2012-06-01"),
                      stringsAsFactors = FALSE)
 
-es <- eventstudy(firm.returns = AMMData[, c("Infosys", "TCS")],
+ammdata <- merge.zoo(Infosys = StockPriceReturns$Infosys,
+                     TCS = StockPriceReturns$TCS,
+                     NiftyIndex,
+                     INRUSD = OtherReturns$INRUSD,
+                     CallMoneyRate = OtherReturns$CallMoneyRate,
+                     all = FALSE)
+ammdata <- window(ammdata, start = "2012-02-01", end = "2012-12-31")
+
+es <- eventstudy(firm.returns = ammdata[, c("Infosys", "TCS")],
                  eventList = events,
                  width = 10,
                  type = "lmAMM",
@@ -231,9 +240,9 @@
                  remap = "cumsum",
                  inference = TRUE,
                  inference.strategy = "bootstrap",
-                 ## model args
-                 market.returns = AMMData[, "index.nifty"],
-                 others = AMMData[, c("currency.inrusd", "call.money.rate")],
+                                                 # model arguments
+                 market.returns = ammdata[, "NiftyIndex"],
+                 others = ammdata[, c("INRUSD", "CallMoneyRate")],
                  market.returns.purge = TRUE
                  )
 str(es)

Modified: pkg/man/excessReturn.Rd
===================================================================
--- pkg/man/excessReturn.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/excessReturn.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -25,15 +25,15 @@
 
 \examples{ 
 data(StockPriceReturns)
-data(NiftyIndex)
+data(OtherReturns)
 
 er.result <- excessReturn(firm.returns = StockPriceReturns,
-			  market.returns = NiftyIndex)
+			  market.returns = OtherReturns$NiftyIndex)
 
-## Checking output: Comparing excess return, raw returns, nifty returns
-output <- merge(er.result$Infosys, StockPriceReturns$Infosys, NiftyIndex,all=FALSE)
-colnames(output) <- c("excess.return", "raw.returns", "nifty.returns")
-tail(output) 
+tail(merge(excessReturn = er.result$Infosys,
+           Infosys = StockPriceReturns$Infosys,
+           NiftyIndex = OtherReturns$NiftyIndex,
+           all=FALSE))
 }
 
 \keyword{excessReturn}

Modified: pkg/man/lmAMM.Rd
===================================================================
--- pkg/man/lmAMM.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/lmAMM.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -77,32 +77,31 @@
 }
 
 \examples{
-data("AMMData")
 data("StockPriceReturns")
-data("NiftyIndex")
+data("OtherReturns")
 
-firm.returns <- AMMData[,"Infosys"]
-market.returns <- AMMData[,"index.nifty"]
-currency.returns <- AMMData[,"currency.inrusd"]
+firm.returns <- StockPriceReturns[, "Infosys"]
+market.returns <- OtherReturns[ ,"NiftyIndex"]
+currency.returns <- OtherReturns[, "INRUSD"]
 
 X <- makeX(market.returns,
-          others = currency.returns,
-          switch.to.innov = FALSE,
-          market.returns.purge = FALSE,
-          verbose = FALSE)
+           others = currency.returns,
+           switch.to.innov = FALSE,
+           market.returns.purge = FALSE,
+           verbose = FALSE)
 
 amm.result <- lmAMM(firm.returns, X, nlags = 0, verbose = FALSE)
 plot(amm.result)
 
 amm.residual <- residuals(amm.result)
-amm.residual <- zoo(amm.residual,as.Date(names(amm.residual)))
+amm.residual <- zoo(amm.residual,
+                    order.by = as.Date(names(amm.residual)))
 
-## Checking output: Comparing augmented market model residual, raw returns, nifty returns
-output <- merge(amm.residual, StockPriceReturns$Infosys, NiftyIndex,
-                all = FALSE)
-colnames(output) <- c("AMM Residual", "Raw Returns", "Nifty Returns")
-tail(output)
-plot(output)
+comparison <- merge(AMMResidual = amm.residual,
+                    Infosys = StockPriceReturns$Infosys,
+                    NiftyIndex =  OtherReturns$NiftyIndex,
+                    all = FALSE)
+plot(comparison)
 }
 
 \keyword{lmAMM}

Modified: pkg/man/makeX.Rd
===================================================================
--- pkg/man/makeX.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/makeX.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -75,15 +75,15 @@
 \author{Ajay Shah, Chirag Anand, Vikram Bahure, Vimal Balasubramaniam}
 
 \examples{
-data("AMMData")
-market.returns <- AMMData$index.nifty
-currency.returns <- AMMData$currency.inrusd
+data("OtherReturns")
+market.returns <- OtherReturns$NiftyIndex
+currency.returns <- OtherReturns$INRUSD
 
 X <- makeX(market.returns,
            others = currency.returns,
            switch.to.innov = FALSE,
            market.returns.purge = FALSE,
            verbose = FALSE)
-head(X)
+head(na.omit(X))
 }
 \keyword{makeX}

Modified: pkg/man/manyfirmssubperiod.lmAMM.Rd
===================================================================
--- pkg/man/manyfirmssubperiod.lmAMM.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/manyfirmssubperiod.lmAMM.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -59,11 +59,11 @@
 }
 
 \examples{
-data("AMMData")
+data("OtherReturns")
 
-firm.returns <- AMMData[, c("Infosys","TCS")]
-market.returns <- AMMData[, "index.nifty"]
-currency.returns <- AMMData[, "currency.inrusd"]
+firm.returns <- StockPriceReturns[, c("Infosys","TCS")]
+market.returns <- OtherReturns$NiftyIndex
+currency.returns <- OtherReturns$INRUSD
 
 X <- makeX(market.returns,
            others = currency.returns,

Modified: pkg/man/marketResidual.Rd
===================================================================
--- pkg/man/marketResidual.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/marketResidual.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -21,18 +21,17 @@
 \author{Vikram Bahure}
 
 \examples{ 
-data(StockPriceReturns)
-data(NiftyIndex)
+data("StockPriceReturns")
+data("OtherReturns")
 
 mm.result <- marketResidual(firm.returns = StockPriceReturns,
-                            market.returns = NiftyIndex)
+                            market.returns = OtherReturns$NiftyIndex)
 
-## Checking output: Comparing market model residual, raw returns, nifty returns
-output <- merge(mm.result$Infosys, StockPriceReturns$Infosys, NiftyIndex, 
-                all=FALSE)
-colnames(output) <- c("market.residual", "raw.returns", "nifty.returns")
-tail(output) 
-
+comparison <- merge(MarketResidual = mm.result$Infosys,
+                    Infosys = StockPriceReturns$Infosys,
+                    NiftyIndex = OtherReturns$NiftyIndex,
+                    all = FALSE)
+plot(comparison)
 }
 
 \keyword{marketResidual}

Modified: pkg/man/phys2eventtime.Rd
===================================================================
--- pkg/man/phys2eventtime.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/phys2eventtime.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -83,15 +83,5 @@
 			 width = 5)
 print(result$z.e[as.character(-3:3)])
 print(result$outcomes)
-
-## Checking conversion to event time frame for first successful event date
-c.no <- as.numeric(colnames(result$z.e))
-cnames <- SplitDates[c.no[1], ]
-phys.output <- as.numeric(result$z.e[as.character(c(-5:5)), as.character(c.no[1])])
-loc <- which(index(StockPriceReturns) \%in\% SplitDates[c.no[1], "event.when"])
-raw.data <- as.numeric(StockPriceReturns[c((loc-5):(loc+5)),
-                       SplitDates[c.no[1], "outcome.unit"]])
-check.output <- cbind(raw.data, phys.output)
-check.output
 }
 \keyword{phys2eventime}

Modified: pkg/man/remap.cumprod.Rd
===================================================================
--- pkg/man/remap.cumprod.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/remap.cumprod.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -47,9 +47,6 @@
                            is.returns = TRUE,
                            base = 100)
 
-check.output <- cbind(es.w[,1], eventtime[,1])
-colnames(check.output) <- c("abnormal.returns", "cumulative.abnormal.returns")
-check.output
-head(eventtime[,1:5])
+print(eventtime[as.character(-3:3), ])
 }
 

Modified: pkg/man/remap.cumsum.Rd
===================================================================
--- pkg/man/remap.cumsum.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/remap.cumsum.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -45,9 +45,5 @@
 es.w <- window(es.results$z.e, start = -5, end = +5)
 eventtime <- remap.cumsum(es.w, is.pc = FALSE, base = 0)
 
-## Comparing abnormal returns (AR) and cumulative abnormal returns (CAR)
-check.output <- cbind(es.w[,1], eventtime[,1])
-colnames(check.output) <- c("abnormal.returns", "cumulative.abnormal.returns")
-check.output
-head(eventtime[,1:5])
+print(eventtime[as.character(-3:3), ])
 }

Modified: pkg/man/remap.event.reindex.Rd
===================================================================
--- pkg/man/remap.event.reindex.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/remap.event.reindex.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -36,5 +36,5 @@
 
 eventtime <- remap.event.reindex(es.w)
 
-head(eventtime[, 1:5])
+eventtime[as.character(-3:3), ]
 }

Modified: pkg/man/subperiod.lmAMM.Rd
===================================================================
--- pkg/man/subperiod.lmAMM.Rd	2014-05-12 10:49:17 UTC (rev 318)
+++ pkg/man/subperiod.lmAMM.Rd	2014-05-12 10:50:15 UTC (rev 319)
@@ -64,12 +64,13 @@
 
 \seealso{ \code{\link{lmAMM}}}
 
-\examples{ 
-data("AMMData")
+\examples{
+data("StockPriceReturns")
+data("OtherReturns")
 
-firm.returns <- AMMData$Infosys
-market.returns <- AMMData$index.nifty
-currency.returns <- AMMData$currency.inrusd
+firm.returns <- StockPriceReturns$Infosys
+market.returns <- OtherReturns$NiftyIndex
+currency.returns <- OtherReturns$USDINR
 
 regressors <- makeX(market.returns,
                     others = currency.returns,

Modified: todo.org
===================================================================
--- todo.org	2014-05-12 10:49:17 UTC (rev 318)
+++ todo.org	2014-05-12 10:50:15 UTC (rev 319)
@@ -41,3 +41,259 @@
 * Testing
   - manual calculation of numbers in the tests
   - revert old tests?
+* plot.amm
+  - Fix the x-axis tick labels: the number is too small
+  - Increase number of plots (the funky way)
+
+* plot.es
+  - "Event study plot capabilities" email on 30th April.
+
+* Ajay's comments
+** On the eesPlot code
+   data.frmt2 <- data.use[which(data.use$cluster.pattern != 0), ]
+
+   Can we please have better variable names.
+
+   hilo1 <- c(-big, big)
+   plot.es.graph.both(es.good.normal, es.bad.normal, es.good.purged,
+   es.bad.purged, width, titlestring, ylab)
+
+   Can we please have better names than hilo1. And, you are making it and
+   not using it.
+
+** Feedback on eesPlot
+   Why do we have eesPlot?
+
+   When I look at the name, I think "Okay, this is a plot function, and
+   why is this not just an S3 plot method". When I see the first one line
+   description on the man page my opinion is confirmed.
+   
+   Then I look deeper and it is absolutely not a plot function! It is a
+   function which figures out a list of events, then runs an event study,
+   and then does a customised plot.
+   
+   We should not have such functions.
+   
+   We should ask the user to run ees() and then run eventstudy() and then
+   use the plot method.
+   
+   Perhaps we should ask the user to do:
+   
+   es.lefttail <- eventstudies(left tail)
+   es.righttail <- eventstudies(right tail)
+   plot(mfrow=c(2,1))
+   plot(es.lefttail, type="blah")
+   plot(es.righttail, type="blah")
+   
+   On an unrelated note, I found it disturbing that the code for
+   eesPlot() does not use ees(). This violates the principle of code
+   reuse. Perhaps we should have the framework where x<-ees() just makes
+   lists of interesting events and then summary(x) generates all those
+   descriptive tables about number of events and run length and so on.
+   
+   Why is the example saying "  ## Generating event study plots (using
+   modified event study methodology)". It looks gauche.
+   
+   There is one spelling mistake in the man page but I've forgotten where
+   it is.
+
+** Feedback on eventstudies::ees
+   1. The entire concept of what we're doing is critically connected
+      to the choice of the event window!!!
+
+   The function and the documentation of the function is silent about
+   this and that's completely wrong.
+
+   Our concept of what's a clean unclustered event is : clean within a
+   stated event window. We never say this. And, it's bad software
+   engineering to hardcode this to a number. This must be an argument to
+   the function.
+
+   2. The title of the function and the first para of the function are
+   quite lame. They say:
+   
+   "This function generates summary statistics for identification and
+   analysis of extreme events.". This mostly leaves me in the dark
+   about what's going on.
+
+   "Tail (Rare) events are often the object of interest in finance.
+   These events are defined as those that have a low probability of
+   occurrence. This function identifies such events based on
+   prob.value mentioned by the user and generates summary
+   statistics about the events. If ‘prob.value’ is 2.5%, events
+   below
+   2.5% (lower tail) and above 97.5% (upper tail) of the
+   distribution
+   are identified as extreme events." This makes the function seem
+   like a massive waste of time. Using R we can trivially find the
+   upper tail observations - no new function is required here. If I
+   read this paragraph I would completely lose interest in the
+   package; I would think these lame developers are taking trivial
+   one/two lines of R code and encoding it as a function with a new
+   name - why would I never bother to learn their new API.
+   
+   The entire value added of the code lies in identifying clean
+   unclustered events, stabbing into messy situations by trying to fuse
+   clustered events under certain conditions, and walking away from
+   places where fusing can't be done. None of that is advertised in the
+   man page. The word 'fuse' does not occur anywhere on the man page!
+   
+   3. When I run the example I get a huge messy structure that's no
+   fun. Why not have: 
+   str(output, max.level=2)
+   which is more comprehensible.
+
+   4. Look at
+   
+      library(eventstudies)
+   data(EESData)
+   ## Input S&P 500 as the univariate series
+   input <- EESData$sp500
+   ## Constructing summary statistics for 5% tail values (5% on both
+   sides)
+   output <- ees(input, prob.value = 5)
+   str(output)
+   
+   It looks nicer and more readable as:
+   
+   library(eventstudies)
+   data(EESData)
+   r <- ees(EESData$sp500, prob.value = 5)
+   str(r, max.level=2)
+
+   5. Choose a consistent style. Is there going to be a
+      library(eventstudies) in front of all the examples? This was not
+      there with the others. Why is it here?
+   
+   6. Why are we saying "   To convert number to words, code uses
+      function “numbers2words†by
+      John Fox and “deprintize†function by Miron Kursa.". We are
+      using thousands of functions by others but is this a big deal?
+   
+   7. In
+   
+      $data$Clustered
+      event.series cluster.pattern
+      2000-03-16     2.524452               3
+      2003-03-17     3.904668               2
+
+      Perhaps the word `runlength' is universally understood instead of
+      cluster.pattern
+
+      The word `event.series' is incomprehensible to me.
+
+   8. In : 
+   
+      > output$upper.tail$extreme.event.distribution
+      unclstr used.clstr removed.clstr tot.clstr tot tot.used
+      upper      65          5            32        37 102       70
+
+      The column names are horrible.
+
+      Pick a more rational sequencing where this process unfolds from
+      left to right.
+
+      This table is the heart of the functionality of what's being done and
+      it isn't explained at all in the man page.
+
+      The man page should say that the researcher might like to only
+      study clean unclustered events - in which case he should run with
+      xxx. If he wishes to use the methodology of fusing adjacent events
+      as done in PSS, then additionally we are able to salvage the events
+      xxx.
+
+
+   9. The run length table should be defined as a table showing a
+      column which is the run length and a column which is the number
+      of events which are a run of that length.
+
+   10. Just confirming: In a package vignette we're going to be able
+       to reproduce some key results from the tables of PSS using this
+       function?
+    
+   11. Wouldn't it be neat to draw something graphical with
+       abline(v=xxx, lty=2) where all the extreme events are shown on
+       a picture? With a different colour for fused and for rejected
+       events.
+
+** Feedback on eventstudies package
+
+   First batch.
+
+   - At many places the phrase `eventstudy' is being used when what's
+     required is `event study'.
+
+   - When I say ?AMMData iqt is riddled with mistakes!!!! The man page
+     has four sentences and has more than 1 error per sentence.
+
+  1. The first few words read: "The data series is a daily time-series
+     zoo object. The sample range for the data is from 2012-02-01 to
+     2014-01-31." Why should this be the top priority?
+
+  2. The two sentences after this, which add up to the full man page,
+     contain one grammatical error each.
+
+  3. Nowhere in the man page is the unit mentioned (per cent).
+
+  4. The dataset contains call.money.rate and that's inconsistent with
+     the man page.
+
+  5. The example says library(zoo) which is not required.
+
+Why do we need a special data object named AMMData? Can we not just
+have one single example dataset with daily returns data for firms,
+that is used for the examples involving both event studies and AMM?
+
+If you had to have this in the package (which I doubt), a better
+example is:
+
+  data(AMMData)
+  str(AMMData)
+  tail(AMMData)
+  summary(AMMData)
+
+We in India use too many abbreviations. Let's stick to the phrase
+`augmented market model' instead of overusing the phrase AMM.
+
+
+*** When I say ?EESData I see a section `Format' which is not in ?AMMData.
+
+    The facts on this man page should say that this is a dataset for
+    the purpose of demonstrating the EES functionality (no
+    abbreviations please), and for replicating the results of the PSS
+    paper. It should explain what the data is (daily returns measured
+    in per cent).
+
+    - Why is the example here different from the example for AMMData?
+
+*** The dataset INR introduces a new word `sample' which was not used in the previous two.
+    Can we please have extreme maniacal consistency in all these?
+    As pointed out above, there is duplication between INR being here and
+    it being in AMMData.
+
+*** It is truly wrong to have a MMData data object!!
+    Nothing prevents you from estimating an MM using the data for an AMM.
+    Can we please be more intelligent about all this.
+
+** Collated
+   - bad variable names
+   - eesPlot: make it S3 function
+     - Do: ees(), eventstudy(), plot()
+   - summary.ees()
+   - ees(): event window in the API and the man pages (language + information)
+   - Remove comments from examples, plus cleaning
+   - Example consistency: remove library() calls from examples
+   - Remove unneeded references
+   - ees(): output colnames, output table format (+sequencing)
+   - ees(): reproducibility of PSS in the vignette
+   - plot.ees()
+   - Spell check
+   - Use "event study" instead of "eventstudy"
+   - Man pages: AMMData: grammatical errors, language, units,
+     consistent sections, call.money.rate
+   - EESData: say about PSS
+   - Avoid abbreviations
+   - Get rid of MMData, INR dataset
+   - lmAMM example
+   - phys2eventtime example
+   - Spell check



More information about the Eventstudies-commits mailing list