Graphing observations over time

[Ian Mac]

Hi,

Sorry if I'm being stupid here (I am a bit of a Stata novice!), but looking for a way to graph number of observations based on a date variable, and filtering using other variables. In particular I want to see the cumulative number of observations over a period of days (of the order of 50 days, with each observation only recorded as a date), with a single point plotted on a graph for each of those days. I also need to be able to filter this data based on other variables at will.

The former I haven't had a problem with as I can create temporary cumulative variables and graph these, but this is a bit of a workaround, and I don't particularly fancy coding this separately for each of the combinations of filters I might wish to apply.

For example (and not what I am working on):

Lets say my observations are purchases of tins of paint. Each has a date associated with the purchase, and is a tin of a specified colour and a specified size. I want to be able to graph the cumulative number of purchases of paint tins over time (with date as my x axis, cumulative number of tins as my y axis), but I also want to be able to easily filter this to look at the same graph for individual colours or size of tins. In my own work the number of combinations of 'colours' or 'sizes' of tins is well into four digits, so coding each separately is far from ideal. In an ideal world these is a single line of code that I can just add 'if colour == XXX & size == YYY' to at will.

Am I just being stupid?

Thanks!
I

[Marcos Almeida]

Welcome to the Stata Forum / Statalist.

I'm not sure if I fully understand your query. Indeed, the best approach - as recommended in the FAQ - is presenting data as an workable example. You may use just a toy example, if you will. Please make sure to share it under CODE delimiters or by installing the SSC - dataex - also recommended in the FAQ.

These comments being made, and considering there is no example with data to work on, I gather you may use time-series analysis for that. For examples - tsset - followed by - tsine - , perhaps with "if condition" as you wish, may do the trick for you.

[Ian Mac]

Hi,

Example data as described above:

* Example generated by -dataex-. To install: ssc install dataex
clear
input float id str5 colour str7 size float purchasedate
  1 "Blue"  "Large"   21058
  2 "Blue"  "X-Large" 21066
  3 "Red"   "Small"   21054
  4 "Red"   "Medium"  21061
  5 "White" "Medium"  21062
  6 "Blue"  "Large"   21066
  7 "Red"   "Small"   21063
  8 "Blue"  "Large"   21065
  9 "Green" "X-Large" 21061
 10 "White" "Small"   21058
 11 "Red"   "Small"   21070
 12 "White" "Medium"  21058
 13 "Green" "X-Large" 21059
 14 "Blue"  "Large"   21057
 15 "White" "Medium"  21059
 16 "Blue"  "X-Large" 21060
 17 "Red"   "X-Large" 21064
 18 "White" "X-Large" 21057
 19 "Green" "X-Large" 21063
 20 "Green" "Small"   21061
 21 "White" "Large"   21057
 22 "Blue"  "Medium"  21057
 23 "Red"   "Large"   21059
 24 "Blue"  "Medium"  21056
 25 "White" "X-Large" 21067
 26 "Green" "Small"   21055
 27 "Red"   "Large"   21064
 28 "Green" "Large"   21061
 29 "Blue"  "Small"   21061
 30 "Green" "X-Large" 21064
 31 "White" "Medium"  21059
 32 "Red"   "Large"   21057
 33 "Blue"  "Medium"  21060
 34 "White" "X-Large" 21059
 35 "White" "Medium"  21054
 36 "Blue"  "Large"   21060
 37 "Red"   "Medium"  21060
 38 "Green" "Large"   21060
 39 "White" "Medium"  21055
 40 "Green" "X-Large" 21057
 41 "White" "Large"   21058
 42 "Blue"  "Medium"  21062
 43 "White" "Large"   21054
 44 "White" "Large"   21062
 45 "Green" "Medium"  21062
 46 "White" "X-Large" 21059
 47 "Green" "Large"   21061
 48 "Red"   "Large"   21061
 49 "White" "X-Large" 21058
 50 "Red"   "Large"   21064
 51 "White" "X-Large" 21056
 52 "Blue"  "Large"   21056
 53 "White" "Medium"  21057
 54 "Green" "Small"   21058
 55 "Blue"  "Large"   21056
 56 "Red"   "Medium"  21061
 57 "White" "Medium"  21061
 58 "Blue"  "Small"   21062
 59 "Red"   "X-Large" 21061
 60 "Green" "Large"   21063
 61 "Green" "X-Large" 21065
 62 "White" "X-Large" 21056
 63 "Green" "Medium"  21061
 64 "Blue"  "Large"   21062
 65 "Green" "Medium"  21062
 66 "Red"   "X-Large" 21058
 67 "Blue"  "Large"   21059
 68 "Red"   "Large"   21062
 69 "Green" "Large"   21061
 70 "Blue"  "X-Large" 21059
 71 "White" "Medium"  21068
 72 "Green" "X-Large" 21060
 73 "Red"   "X-Large" 21060
 74 "Blue"  "Medium"  21060
 75 "Red"   "Small"   21060
 76 "Green" "X-Large" 21060
 77 "Blue"  "Large"   21066
 78 "Red"   "Small"   21060
 79 "Red"   "Large"   21068
 80 "Red"   "X-Large" 21057
 81 "Blue"  "Medium"  21063
 82 "Blue"  "Small"   21061
 83 "Red"   "Large"   21067
 84 "Red"   "Medium"  21065
 85 "Red"   "X-Large" 21066
 86 "Blue"  "Medium"  21061
 87 "Red"   "X-Large" 21056
 88 "Red"   "X-Large" 21060
 89 "White" "X-Large" 21060
 90 "White" "X-Large" 21059
 91 "Red"   "Medium"  21064
 92 "White" "Medium"  21057
 93 "White" "X-Large" 21057
 94 "White" "Small"   21060
 95 "Red"   "Medium"  21063
 96 "Green" "Medium"  21059
 97 "Blue"  "X-Large" 21057
 98 "Blue"  "Medium"  21060
 99 "Green" "Large"   21057
100 "Blue"  "X-Large" 21061
end
format %td purchasedate

My understanding (which is, admittedly, quite limited) is that Time Series wouldn't be appropriate here as the observations are the individual purchases (and therefore the same date will occur within many observations), rather than the dates. Although I guess if there is a way to change to represent the data this way then this would be possible!

Thanks,
I

[Nick Cox]

This sounds just like graphs of the cumulative frequencies of your date variable, qualified as you wish by colour, size or any combination thereof.

distplot (Stata Journal) is a dedicated command to install. Of the entries below, the latest indicates which software you can see -- indeed as I write a further cosmetic update is in press in Stata Journal 17(3).

. search distplot , sj historical

Search of official help files, FAQs, Examples, SJs, and STBs

SJ-10-1 gr41_4  . . . . . . . . . . . . . . . . . Software update for distplot
        (help distplot if installed)  . . . . . . . . . . . . . . .  N. J. Cox
        Q1/10   SJ 10(1):164
        new reverse(ge) option specifies plotting probabilities or
        frequencies greater than or equal to any data value

SJ-5-3  gr0018  . . . . . . . . . .  Speaking Stata: The protean quantile plot
        . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  N. J. Cox
        Q3/05   SJ 5(3):442--460           (see gr41_3 and gr42_3 for commands)
        discusses quantile and distribution plots as used in
        the analysis of species abundance data in ecology

SJ-5-3  gr41_3  . . . . . . . . . . . . . . . . . Software update for distplot
        (help distplot if installed)  . . . . . . . . . . . . . . .  N. J. Cox
        Q3/05   SJ 5(3):471
        simplified syntax; both by() and over() are now allowed

SJ-4-2  gr0004  .  Speaking Stata: Graphing categorical and compositional data
        . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  N. J. Cox
        Q2/04   SJ 4(2):190--215                                 (no commands)
        discusses graphical possibilities for categorical and
        compositional data

SJ-4-1  gr0003  . . . . . . . . . . . . Speaking Stata: Graphing distributions
        . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .  N. J. Cox
        Q1/04   SJ 4(1):66--88                                   (no commands)
        a review of official and user-written commands for
        graphing univariate distributions; includes tricks
        beyond what is obviously and readily available

SJ-3-4  gr41_2  . . . . . . . . . . . . . . . . . Software update for distplot
        (help distplot if installed)  . . . . . . . . . . . . . . .  N. J. Cox
        Q4/03   SJ 3(4):449
        option tscale() renamed as trscale()

SJ-3-2  gr41_1  . . . . . . . . . . . . . . . . . Software update for distplot
        (help distplot if installed)  . . . . . . . . . . . . . . .  N. J. Cox
        Q2/03   SJ 3(2):211
        enhanced to use Stata 8 graphics and provides new options

STB-51  gr41  . . . . . . . . . . . . . . . . . .  Distribution function plots
        (help distplot if installed)  . . . . . . . . . . . . . . .  N. J. Cox
        9/99    pp.12--16; STB Reprints Vol 9, pp.108--112
        plots the cumulative distribution function or survival function
        and allows multiple variables

Here are some token code examples:

distplot purchasedate, over(colour) frequency c(J..)

distplot purchasedate if colour == "Blue", frequency c(J)

Please note: We ask for full real names here, not nicknames, abbreviations, or pseudonyms. The FAQ Advice explains further.

[Ian Mac]

That's excellent, thank you! I knew there must have been an easier way!

I

[Nick Cox]

Correction to #4: An update to distplot will not be appearing in SJ 17(3). I confused intention with action. I added a feature, or fixed a misfeature, but have yet to finish the task.

[Martin Wissing]

Dear Nick, thank you for the distplot program. This is really a basic graph which has been missing in STATA.

[Nick Cox]

#7 Thanks for the thanks. The StataCorp line (so to speak) on this can be seen at

help cumul

which underlines that cumul can calculate the (empirical (cumulative)) distribution function and line can then plot it.