Plotting multiple annual time series over top each other

I have a dataframe containing daily observations of various climate measurements for 31 stations, which are factors. Each station has many years' worth of daily observations and effectively, each station has a unique number of years recorded, and unique number of observations.

For example data, I have subset it down to a 13 stations with one observation per unique water_year.

NAME DATE PRCP calendar_year month day water_year water_date <fct> <date> <dbl> <fct> <int> <int> <fct> <date> 102 FLORENCE 0.2 SSE, OR US 2007-12-05 0 2007 12 5 2007 2006-12-05 103 FLORENCE 0.2 SSE, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 104 FLORENCE 0.2 SSE, OR US 2009-12-16 0.9 2009 12 16 2009 2008-12-16 105 FLORENCE 0.2 SSE, OR US 2010-10-19 0 2010 10 19 2010 2009-10-19 106 FLORENCE 0.2 SSE, OR US 2012-07-10 0 2012 7 10 2012 2012-07-10 107 FLORENCE 0.5 NE, OR US 2007-12-12 0 2007 12 12 2007 2006-12-12 108 FLORENCE 0.5 NE, OR US 2008-01-01 0 2008 1 1 2008 2008-01-01 109 FLORENCE 0.6 E, OR US 2008-01-01 0 2008 1 1 2008 2008-01-01 110 FLORENCE 0.9 NW, OR US 2007-12-22 0.09 2007 12 22 2007 2006-12-22 111 FLORENCE 0.9 NW, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 112 FLORENCE 0.9 NW, OR US 2009-10-01 0.02 2009 10 1 2009 2008-10-01 113 FLORENCE 0.9 NW, OR US 2010-10-01 0.03 2010 10 1 2010 2009-10-01 114 FLORENCE 0.9 NW, OR US 2011-10-01 0.02 2011 10 1 2011 2010-10-01 115 FLORENCE 0.9 NW, OR US 2012-10-01 0 2012 10 1 2012 2011-10-01 116 FLORENCE 0.9 NW, OR US 2013-10-01 0.92 2013 10 1 2013 2012-10-01 117 FLORENCE 0.9 NW, OR US 2014-10-01 0.01 2014 10 1 2014 2013-10-01 118 FLORENCE 0.9 NW, OR US 2015-10-01 0 2015 10 1 2015 2014-10-01 119 FLORENCE 0.9 NW, OR US 2016-10-01 0.15 2016 10 1 2016 2015-10-01 120 FLORENCE 0.9 NW, OR US 2017-10-01 0.2 2017 10 1 2017 2016-10-01 121 FLORENCE 0.9 NW, OR US 2018-01-01 0 2018 1 1 2018 2018-01-01 122 FLORENCE 1.8 NW, OR US 2007-12-14 0 2007 12 14 2007 2006-12-14 123 FLORENCE 1.8 NW, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 124 FLORENCE 1.8 NW, OR US 2009-10-25 0 2009 10 25 2009 2008-10-25 125 FLORENCE 1.8 NW, OR US 2010-10-05 0.01 2010 10 5 2010 2009-10-05 126 FLORENCE 1.8 NW, OR US 2011-10-01 0.02 2011 10 1 2011 2010-10-01 127 FLORENCE 1.8 NW, OR US 2012-10-02 0 2012 10 2 2012 2011-10-02 128 FLORENCE 1.8 NW, OR US 2013-10-01 0.570 2013 10 1 2013 2012-10-01 129 FLORENCE 1.8 NW, OR US 2014-10-01 0.02 2014 10 1 2014 2013-10-01 130 FLORENCE 1.8 NW, OR US 2015-10-01 0.02 2015 10 1 2015 2014-10-01 131 FLORENCE 1.8 NW, OR US 2016-10-01 0.08 2016 10 1 2016 2015-10-01 132 FLORENCE 1.8 NW, OR US 2017-10-01 0.23 2017 10 1 2017 2016-10-01 133 FLORENCE 1.8 NW, OR US 2018-01-01 0.01 2018 1 1 2018 2018-01-01 134 FLORENCE 2.1 NNW, OR US 2007-12-17 0.96 2007 12 17 2007 2006-12-17 135 FLORENCE 2.1 NNW, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 136 FLORENCE 2.1 NNW, OR US 2009-10-01 0 2009 10 1 2009 2008-10-01 137 FLORENCE 2.1 NNW, OR US 2010-10-01 0.03 2010 10 1 2010 2009-10-01 138 FLORENCE 2.1 NNW, OR US 2011-10-01 0 2011 10 1 2011 2010-10-01 139 FLORENCE 2.1 NNW, OR US 2012-10-01 0 2012 10 1 2012 2011-10-01 140 FLORENCE 2.1 NNW, OR US 2013-12-26 0 2013 12 26 2013 2012-12-26 141 FLORENCE 2.1 NNW, OR US 2014-10-07 0 2014 10 7 2014 2013-10-07 142 FLORENCE 2.1 NNW, OR US 2016-05-21 0 2016 5 21 2016 2016-05-21 143 FLORENCE 2.1 NNW, OR US 2017-12-26 0 2017 12 26 2017 2016-12-26 144 FLORENCE 2.9 NNW, OR US 2007-12-16 0.07 2007 12 16 2007 2006-12-16 145 FLORENCE 2.9 NNW, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 146 FLORENCE 2.9 NNW, OR US 2009-10-01 0.03 2009 10 1 2009 2008-10-01 147 FLORENCE 2.9 NNW, OR US 2010-10-02 0.05 2010 10 2 2010 2009-10-02 148 FLORENCE 2.9 NNW, OR US 2011-10-01 0.02 2011 10 1 2011 2010-10-01 149 FLORENCE 2.9 NNW, OR US 2012-10-02 0 2012 10 2 2012 2011-10-02 150 FLORENCE 2.9 NNW, OR US 2013-10-01 0.580 2013 10 1 2013 2012-10-01 151 FLORENCE 2.9 NNW, OR US 2014-10-01 0.02 2014 10 1 2014 2013-10-01 152 FLORENCE 2.9 NNW, OR US 2015-10-01 0 2015 10 1 2015 2014-10-01 153 FLORENCE 2.9 NNW, OR US 2016-10-04 0.580 2016 10 4 2016 2015-10-04 154 FLORENCE 2.9 NNW, OR US 2017-10-01 0.2 2017 10 1 2017 2016-10-01 155 FLORENCE 2.9 NNW, OR US 2018-01-01 0 2018 1 1 2018 2018-01-01 156 FLORENCE 5.4 N, OR US 2007-12-22 0.03 2007 12 22 2007 2006-12-22 157 FLORENCE 5.4 N, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 158 FLORENCE 5.4 N, OR US 2009-10-01 0.07 2009 10 1 2009 2008-10-01 159 FLORENCE 5.4 N, OR US 2010-10-01 0.03 2010 10 1 2010 2009-10-01 160 FLORENCE 5.4 N, OR US 2011-10-03 0.65 2011 10 3 2011 2010-10-03 161 FLORENCE 5.4 N, OR US 2012-10-01 0 2012 10 1 2012 2011-10-01 162 FLORENCE 5.4 N, OR US 2013-10-01 0.6 2013 10 1 2013 2012-10-01 163 FLORENCE 5.4 N, OR US 2014-10-01 0 2014 10 1 2014 2013-10-01 164 FLORENCE 5.4 N, OR US 2015-10-01 0 2015 10 1 2015 2014-10-01 165 FLORENCE 5.4 N, OR US 2016-11-01 0.21 2016 11 1 2016 2015-11-01 166 FLORENCE 5.4 N, OR US 2017-11-11 0.9 2017 11 11 2017 2016-11-11 167 FLORENCE 5.4 N, OR US 2018-01-01 0 2018 1 1 2018 2018-01-01 168 FLORENCE 5.4 S, OR US 2007-12-08 0.42 2007 12 8 2007 2006-12-08 169 FLORENCE 5.4 S, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 170 FLORENCE 5.4 S, OR US 2009-10-01 0 2009 10 1 2009 2008-10-01 171 FLORENCE 5.4 S, OR US 2010-10-01 0.03 2010 10 1 2010 2009-10-01 172 FLORENCE 5.4 S, OR US 2011-10-01 0 2011 10 1 2011 2010-10-01 173 FLORENCE 5.4 S, OR US 2012-10-01 0 2012 10 1 2012 2011-10-01 174 FLORENCE 5.4 S, OR US 2013-10-01 0.6 2013 10 1 2013 2012-10-01 175 FLORENCE 5.4 S, OR US 2014-10-02 0 2014 10 2 2014 2013-10-02 176 FLORENCE 5.4 S, OR US 2015-01-01 0 2015 1 1 2015 2015-01-01 177 FLORENCE 5.8 S, OR US 2007-12-01 0.02 2007 12 1 2007 2006-12-01 178 FLORENCE 5.8 S, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 179 FLORENCE 5.8 S, OR US 2009-10-01 0.02 2009 10 1 2009 2008-10-01 180 FLORENCE 5.8 S, OR US 2010-10-01 0.01 2010 10 1 2010 2009-10-01 181 FLORENCE 5.8 S, OR US 2011-10-01 0 2011 10 1 2011 2010-10-01 182 FLORENCE 5.8 S, OR US 2012-10-01 0 2012 10 1 2012 2011-10-01 183 FLORENCE 5.8 S, OR US 2013-10-01 0.75 2013 10 1 2013 2012-10-01 184 FLORENCE 5.8 S, OR US 2014-01-01 0 2014 1 1 2014 2014-01-01 185 FLORENCE 5.9 NNE, OR US 2007-11-29 0.41 2007 11 29 2007 2006-11-29 186 FLORENCE 5.9 NNE, OR US 2008-10-03 0.39 2008 10 3 2008 2007-10-03 187 FLORENCE 5.9 NNE, OR US 2009-10-01 0.01 2009 10 1 2009 2008-10-01 188 FLORENCE 5.9 NNE, OR US 2010-10-01 0.05 2010 10 1 2010 2009-10-01 189 FLORENCE 5.9 NNE, OR US 2011-10-01 0.02 2011 10 1 2011 2010-10-01 190 FLORENCE 5.9 NNE, OR US 2012-10-01 0 2012 10 1 2012 2011-10-01 191 FLORENCE 5.9 NNE, OR US 2013-10-01 0.43 2013 10 1 2013 2012-10-01 192 FLORENCE 5.9 NNE, OR US 2014-10-01 0 2014 10 1 2014 2013-10-01 193 FLORENCE 5.9 NNE, OR US 2015-10-10 0.69 2015 10 10 2015 2014-10-10 194 FLORENCE 5.9 NNE, OR US 2016-10-01 0.11 2016 10 1 2016 2015-10-01 195 FLORENCE 5.9 NNE, OR US 2017-01-01 0.24 2017 1 1 2017 2017-01-01 196 FLORENCE 6 N, OR US 2007-11-19 0.04 2007 11 19 2007 2006-11-19 197 FLORENCE 6 N, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 198 FLORENCE 6 N, OR US 2009-10-01 0 2009 10 1 2009 2008-10-01 199 FLORENCE 6 N, OR US 2010-01-01 0.7 2010 1 1 2010 2010-01-01 200 FLORENCE NUMBER 2, OR US 2006-10-01 0 2006 10 1 2006 2005-10-01 201 FLORENCE NUMBER 2, OR US 2007-10-01 0 2007 10 1 2007 2006-10-01 202 FLORENCE NUMBER 2, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 203 FLORENCE NUMBER 2, OR US 2009-10-01 0 2009 10 1 2009 2008-10-01 204 FLORENCE NUMBER 2, OR US 2010-10-01 0.04 2010 10 1 2010 2009-10-01 205 FLORENCE NUMBER 2, OR US 2011-10-01 0.9 2011 10 1 2011 2010-10-01 206 FLORENCE NUMBER 2, OR US 2012-10-01 0 2012 10 1 2012 2011-10-01 207 FLORENCE NUMBER 2, OR US 2013-10-01 0.46 2013 10 1 2013 2012-10-01 208 FLORENCE NUMBER 2, OR US 2014-10-01 0 2014 10 1 2014 2013-10-01 209 FLORENCE NUMBER 2, OR US 2015-10-01 0 2015 10 1 2015 2014-10-01 210 FLORENCE NUMBER 2, OR US 2016-10-01 0.77 2016 10 1 2016 2015-10-01 211 FLORENCE NUMBER 2, OR US 2017-10-01 0.06 2017 10 1 2017 2016-10-01 212 FLORENCE NUMBER 2, OR US 2018-01-01 0 2018 1 1 2018 2018-01-01 213 FLORENCE, OR US 1909-10-01 0.580 1909 10 1 1909 1908-10-01 214 FLORENCE, OR US 1910-10-01 0.49 1910 10 1 1910 1909-10-01 215 FLORENCE, OR US 1911-10-01 0.03 1911 10 1 1911 1910-10-01 216 FLORENCE, OR US 1912-10-01 0.07 1912 10 1 1912 1911-10-01 217 FLORENCE, OR US 1913-10-01 0 1913 10 1 1913 1912-10-01 218 FLORENCE, OR US 1914-10-01 0.24 1914 10 1 1914 1913-10-01 219 FLORENCE, OR US 1915-10-01 0.25 1915 10 1 1915 1914-10-01 220 FLORENCE, OR US 1916-10-01 0.03 1916 10 1 1916 1915-10-01 221 FLORENCE, OR US 1917-10-01 0 1917 10 1 1917 1916-10-01 222 FLORENCE, OR US 1918-10-01 0 1918 10 1 1918 1917-10-01 223 FLORENCE, OR US 1919-10-01 0.6 1919 10 1 1919 1918-10-01 224 FLORENCE, OR US 1920-10-01 1.22 1920 10 1 1920 1919-10-01 225 FLORENCE, OR US 1921-10-01 0 1921 10 1 1921 1920-10-01 226 FLORENCE, OR US 1922-10-01 0.03 1922 10 1 1922 1921-10-01 227 FLORENCE, OR US 1949-12-08 0 1949 12 8 1949 1948-12-08 228 FLORENCE, OR US 1950-10-01 0 1950 10 1 1950 1949-10-01 229 FLORENCE, OR US 1951-01-01 0.32 1951 1 1 1951 1951-01-01 230 FLORENCE, OR US 2004-10-01 0 2004 10 1 2004 2003-10-01 231 FLORENCE, OR US 2005-10-01 0.88 2005 10 1 2005 2004-10-01 232 FLORENCE, OR US 2006-10-01 0 2006 10 1 2006 2005-10-01 233 FLORENCE, OR US 2007-10-01 0.33 2007 10 1 2007 2006-10-01 234 FLORENCE, OR US 2008-10-01 0 2008 10 1 2008 2007-10-01 235 FLORENCE, OR US 2009-10-01 0 2009 10 1 2009 2008-10-01 236 FLORENCE, OR US 2010-10-01 0.04 2010 10 1 2010 2009-10-01 237 FLORENCE, OR US 2011-10-01 0.75 2011 10 1 2011 2010-10-01 238 FLORENCE, OR US 2012-10-02 0 2012 10 2 2012 2011-10-02 239 FLORENCE, OR US 2013-10-01 0.63 2013 10 1 2013 2012-10-01 240 FLORENCE, OR US 2014-10-01 0 2014 10 1 2014 2013-10-01 241 FLORENCE, OR US 2015-10-01 0 2015 10 1 2015 2014-10-01 242 FLORENCE, OR US 2016-10-01 0.16 2016 10 1 2016 2015-10-01 243 FLORENCE, OR US 2017-01-01 0.53 2017 1 1 2017 2017-01-01

My goal is to:

So the resulting plots would be PRCP on the y axis, water_date on the x axis, and dots/smooths grouped by each water_year (available for that NAME) plotted over top each other. There would be 31 plots in total, one for each NAME.

A simple way to do this for a given NAME with PRCP plotted against water_date per single water_year would be:

ggplot(srb_clean %>% filter(NAME == "made up name" & water_year == "1902") ,aes(water_date, PRCP)) + geom_point(na.rm=TRUE) + geom_smooth(colour = "red",size = 1)

This code would produce a plot for one year's worth of data whereas the desired out put would have a dot/smooth group for each year of data available for that NAME.

enter image description here

I am looking for a way to automate the process of creating each of these plots, and outputting one plot per NAME, with PRCP x water_date, grouped by water_year.

What is the most elegant, or the most standard way of doing something like this R? I am a programming novice, and somewhat befuddled about how to approach this programmatically, let alone in R in particular.

UPDATE #1 (improved example data and question)

UPDATE #3 (solution)

Parfait's solution works well. It can be used with code similar to that above to output a plots similar to the following:

multiple water years overplot

You should have a closer look at how ggplot work. Especially you want to consider the aes group and use it to group by year, and facet_grid to plot each station. This is basic ggplot, make a bit more research before posting. Also try to implement something before asking for a solution. Good luck ;)
– Hobo Sheep
Aug 25 at 20:19

aes group

facet_grid

This will be a relatively straight-forward task in ggplot -- a reproducible example would help. The toughest part will be getting the x-axis to range from Jan 1 to Dec 31 -- you will have to strip the month + date out of water_date and create a new column with the dates within a dummy year, as done in stackoverflow.com/questions/33832776/… . After that, just feed the dataframe into ggplot, with the aesthetics x = the new dummy date, y = PRCP, grouping with the calendar_year, and facet_wrap by STATION (not facet_grid).
– jhchou
Aug 25 at 20:58

Please share sample of your data using dput() (not str or head or picture/screenshot) so others can help. See more here stackoverflow.com/questions/5963269/…
– Tung
Aug 25 at 21:33

dput()

str

head

Can you also post or draw your desired output plot?
– Tung
Aug 25 at 21:35

@ClaytonGlasser: you need to use dput() to share data. The above table is not readily usable for helpers. The NAME column even contains many spaces
– Tung
Aug 26 at 21:38

dput()

NAME

1 Answer
1

Since you require the same x-axis of dates in an annual period, consider updating all years in water_date to a common year that currently no rows maintain such as 2099 - 2100.

Then use by (the function to slice a dataframe to smaller subsets by one or more factors) to generate a list of plots for each distinct NAME. To ignore the 2099, use the scales library to plot the month and day: %b-%d (month name) %m-%d (month number). Also, pass water_year as fill factor for legend series.

by

scales

%b-%d

%m-%d

library(ggplot2) library(scales) ... # TEMP HELPER VARIABLE df$wt_date_char <- as.character(df$water_date) # REPLACE EVERY YEAR FOR 2099 OR 2100 # CONDITIONALLY UPDATE YEAR BY MONTH NUMEBR df$pseudo_water_date <- ifelse(substr(df$wt_date_char, 6, 7) %in% paste0("0", as.character(seq(1,9))), gsub("^(.*?)\-", "2099-", df$wt_date_char), gsub("^(.*?)\-", "2100-", df$wt_date_char) ) df$pseudo_water_date <- as.Date(df$pseudo_water_date, origin="1970-01-01") df$wt_date_char <- NULL # BUILD PLOT LIST plot_list <- by(srb_clean, srb_clean$NAME, function(sub) ggplot(sub, aes(pseudo_water_date, PRCP, fill=factor(water_year))) + geom_point(na.rm=TRUE) + geom_smooth(colour = "red", size = 1) + ggtitle(sub$NAME[[1]]) + labs(title="Water Year", x="Water Date", y="Precipitation") + theme(plot.title = element_text(hjust = 0.5)) + scale_x_date(labels = date_format("%b-%d")) ) # OUTPUT INDIVIDUAL PLOTS plot_list[[1]] plot_list[[2]] plot_list[[3]] ... # OUTPUT ALL PLOTS plot_list

This is excellent, thank you. This works except in one respect. By converting all the water_dates to year=2099 (enabling me to plot them all on the same annual chart), I lose the ability to make the graph run Oct-Nov (since a full water_year contains 9 months with last year's calendar year, thus "starting" in Oct). How would you suggest obliging the graph to run Oct-Nov (for example 10-01-2000 -- 09-01-2001) with this approach?
– Clayton Glasser
Aug 27 at 4:44

See updated answer adding a conditional ifelse assignment where months 10, 11, 12 are updated to 2099 and all other months, 1-9 are updated to 2100, one year later. For clarity, I use a new variable for graphing purposes: pseudo_water_date.
– Parfait
Aug 27 at 14:46

ifelse

I have been trying to implement this code and have discovered that it has the effect of converting ALL of the years in pseudo_water_date to 2100, not just months 1-9. I have carefully reviewed every aspect of the ifelse/substr/gsub code; it all makes sense and I don't see any errors. Do you have any insight into why it might not work? Is the IF statement not being triggered?
– Clayton Glasser
Aug 29 at 21:00

Whoops! Logic should be adjusted. See edit, changing seq(10,12) to seq(1,9).
– Parfait
Aug 29 at 21:07

seq(10,12)

seq(1,9)

Ah, yes, true. Combined with reversing the gsub logic (switching 2100 and 2099), this has the desired result. Thanks! @Parfait
– Clayton Glasser
Aug 29 at 21:46

By clicking "Post Your Answer", you acknowledge that you have read our updated terms of service, privacy policy and cookie policy, and that your continued use of the website is subject to these policies.

搜尋此網誌

Dfyjkt