handling missingness
play

Handling Missingness Manipulating Time Series Data in R: Case - PowerPoint PPT Presentation

MANIPULATING TIME SERIES DATA IN R: CASE STUDIES Handling Missingness Manipulating Time Series Data in R: Case Studies Missingness > citydata citydata pop 1980-01-01 562994 1981-01-01 564179 1982-01-01 565361 570000 1983-01-01 565491


  1. MANIPULATING TIME SERIES DATA IN R: CASE STUDIES Handling Missingness

  2. Manipulating Time Series Data in R: Case Studies Missingness > citydata citydata pop 1980-01-01 562994 1981-01-01 564179 1982-01-01 565361 570000 1983-01-01 565491 1984-01-01 566723 1985-01-01 NA 564000 1986-01-01 NA 1987-01-01 NA 1988-01-01 570867 Jan Jan Jan Jan Jan Jan 1989-01-01 572222 1980 1982 1984 1986 1988 1990 1990-01-01 574823

  3. Manipulating Time Series Data in R: Case Studies Fill NAs with Last Observation • Last observation carried forward (LOCF) > citydata_locf <- na.locf(citydata) > plot.xts(citydata) > plot.xts(citydata_locf) citydata citydata_locf 570000 570000 564000 564000 Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan 1980 1982 1984 1986 1988 1990 1980 1982 1984 1986 1988 1990

  4. Manipulating Time Series Data in R: Case Studies Fill NAs with Next Observation ● Next observation carried backward (NOCB) > citydata_nocb <- na.locf(citydata, fromLast = TRUE) > plot.xts(citydata) > plot.xts(citydata_nocb) citydata citydata_nocb 570000 570000 564000 564000 Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan 1980 1982 1984 1986 1988 1990 1980 1982 1984 1986 1988 1990

  5. Manipulating Time Series Data in R: Case Studies Linear Interpolation > citydata_approx <- na.approx(citydata) > plot.xts(citydata) > plot.xts(citydata_nocb) citydata citydata_approx 570000 570000 564000 564000 Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan Jan 1980 1982 1984 1986 1988 1990 1980 1982 1984 1986 1988 1990

  6. MANIPULATING TIME SERIES DATA IN R: CASE STUDIES Let’s practice!

  7. MANIPULATING TIME SERIES DATA IN R: CASE STUDIES Lagging and Differencing

  8. Manipulating Time Series Data in R: Case Studies Lagging ● lag() o ff sets observations in time lag(unemployment, k = 1, ...) Jan 2010 9.6 - Feb 2010 9.2 9.6 March 2010 8.9 9.2 April 2010 8.3 8.9 May 2010 8.2 8.3 June 2010 8.4 8.2 July 2010 8.3 8.4

  9. Manipulating Time Series Data in R: Case Studies Differencing • di ff () measures change between periods diff(unemployment, lag = 1, ...) Jan 2010 9.6 - Feb 2010 9.2 -0.4 March 2010 8.9 -0.3 April 2010 8.3 -0.6 May 2010 8.2 -0.1 June 2010 8.4 0.2 July 2010 8.3 -0.1

  10. MANIPULATING TIME SERIES DATA IN R: CASE STUDIES Let’s practice!

  11. MANIPULATING TIME SERIES DATA IN R: CASE STUDIES Rolling Functions

  12. Manipulating Time Series Data in R: Case Studies Discrete Windows ● Split the data according to period > unemployment_yrs <- split(unemployment, f = "years") ● Apply function within period > unemployment_yrs <- lapply(unemployment_yrs, cummax) ● Bind new data into xts object > unemployment_ytd <- do.call(rbind, unemployment_yrs)

  13. Manipulating Time Series Data in R: Case Studies Rolling Windows ● rollapply() applies a function to a rolling window > unemployment_avg <- rollapply(unemployment, width = 12, FUN = mean) Unemployment (%) Unemployment (%) 8 8 6 6 4 4 2 2 2000 2002 2004 2006 2008 2010 2000 2002 2004 2006 2008 2010

  14. MANIPULATING TIME SERIES DATA IN R: CASE STUDIES Let’s practice!

Recommend


More recommend