Line Graph

[Nicholas McPherson]

Hi all,

I am attempting to create a line graph for NO2 levels across four geographical areas (coded as MMM). I am trying to make it so the total levels of NO2 for each geographical area are plotted.

I have tried to use the following basic code: twoway line no2 mmm - however, this produces a very messy graph with multiple lines (see attached).

I then tried to use: twoway line no2 mmm, sorted - however, this too does not work appropriately (see attached).

I am slo trying to include confidence intervals in this graph, however, am just trying to work out how to do this first step!

Any help would be incredibly appreciated!

[Nick Cox]

Why a line graph at all if 1 2 3 4 are areas? Are they in some kind of spatial sequence? What is the overall structure of the data? Please give a data example using dataex.

[Nicholas McPherson]

Why a line graph at all if 1 2 3 4 are areas? Are they in some kind of spatial sequence? What is the overall structure of the data? Please give a data example using dataex.

Hi Nick, my apologies I should have provided more context - 1 to 4 is a scale that denotes rurality of the area in a sequence scale (i.e. 1 = is most urban and 4 = the most remote). Having difficulty giving a data example using dataex (was referring to this thread: https://www.statalist.org/forums/for...-for-statalist)

I have come to realise that I probably need to graph the mean of the NO2 variable so that way I do not have many, many points on the graph etc. is there a code that would allow for this so that I can graph the mean of the NO2 over the MMM groups? (While also including confidence intervals)

[Maarten Buis]

The easiest solution is to realize that linear regression is just computing conditional means, and that is exactly what you want. Here is an example:

. sysuse auto, clear
(1978 automobile data)

. reg price i.rep78

      Source |       SS           df       MS      Number of obs   =        69
-------------+----------------------------------   F(4, 64)        =      0.24
       Model |  8360542.63         4  2090135.66   Prob > F        =    0.9174
    Residual |   568436416        64     8881819   R-squared       =    0.0145
-------------+----------------------------------   Adj R-squared   =   -0.0471
       Total |   576796959        68  8482308.22   Root MSE        =    2980.2

------------------------------------------------------------------------------
       price | Coefficient  Std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       rep78 |
          2  |   1403.125   2356.085     0.60   0.554    -3303.696    6109.946
          3  |   1864.733   2176.458     0.86   0.395    -2483.242    6212.708
          4  |       1507   2221.338     0.68   0.500    -2930.633    5944.633
          5  |     1348.5   2290.927     0.59   0.558    -3228.153    5925.153
             |
       _cons |     4564.5   2107.347     2.17   0.034     354.5913    8774.409
------------------------------------------------------------------------------

.
. // get the means from the model
. margins i.rep78

Adjusted predictions                                        Number of obs = 69
Model VCE: OLS

Expression: Linear prediction, predict()

------------------------------------------------------------------------------
             |            Delta-method
             |     Margin   std. err.      t    P>|t|     [95% conf. interval]
-------------+----------------------------------------------------------------
       rep78 |
          1  |     4564.5   2107.347     2.17   0.034     354.5913    8774.409
          2  |   5967.625   1053.673     5.66   0.000     3862.671    8072.579
          3  |   6429.233   544.1145    11.82   0.000      5342.24    7516.227
          4  |     6071.5   702.4489     8.64   0.000     4668.197    7474.803
          5  |       5913   898.5756     6.58   0.000     4117.889    7708.111
------------------------------------------------------------------------------

.
. // see that these margins are really just means
. table (rep78), stat(mean price)

------------------------------
                   |      Mean
-------------------+----------
Repair record 1978 |          
  1                |    4564.5
  2                |  5967.625
  3                |  6429.233
  4                |    6071.5
  5                |      5913
  Total            |  6146.043
------------------------------

.
. // turn it in the graph we want
. qui margins i.rep78

. marginsplot

[Nicholas McPherson]

The easiest solution is to realize that linear regression is just computing conditional means, and that is exactly what you want. Here is an example:

. sysuse auto, clear
(1978 automobile data)

. reg price i.rep78

Source | SS df MS Number of obs = 69
-------------+---------------------------------- F(4, 64) = 0.24
Model | 8360542.63 4 2090135.66 Prob > F = 0.9174
Residual | 568436416 64 8881819 R-squared = 0.0145
-------------+---------------------------------- Adj R-squared = -0.0471
Total | 576796959 68 8482308.22 Root MSE = 2980.2

------------------------------------------------------------------------------
price | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
rep78 |
2 | 1403.125 2356.085 0.60 0.554 -3303.696 6109.946
3 | 1864.733 2176.458 0.86 0.395 -2483.242 6212.708
4 | 1507 2221.338 0.68 0.500 -2930.633 5944.633
5 | 1348.5 2290.927 0.59 0.558 -3228.153 5925.153
|
_cons | 4564.5 2107.347 2.17 0.034 354.5913 8774.409
------------------------------------------------------------------------------

.
. // get the means from the model
. margins i.rep78

Adjusted predictions Number of obs = 69
Model VCE: OLS

Expression: Linear prediction, predict()

------------------------------------------------------------------------------
| Delta-method
| Margin std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
rep78 |
1 | 4564.5 2107.347 2.17 0.034 354.5913 8774.409
2 | 5967.625 1053.673 5.66 0.000 3862.671 8072.579
3 | 6429.233 544.1145 11.82 0.000 5342.24 7516.227
4 | 6071.5 702.4489 8.64 0.000 4668.197 7474.803
5 | 5913 898.5756 6.58 0.000 4117.889 7708.111
------------------------------------------------------------------------------

.
. // see that these margins are really just means
. table (rep78), stat(mean price)

------------------------------
| Mean
-------------------+----------
Repair record 1978 |
1 | 4564.5
2 | 5967.625
3 | 6429.233
4 | 6071.5
5 | 5913
Total | 6146.043
------------------------------

.
. // turn it in the graph we want
. qui margins i.rep78

. marginsplot

[ATTACH=CONFIG]n1683747[/ATTACH]

Thank you kindly for your response Maarten. Apologies if this is a basic question, but what code should I use to try and get the mean values for my NO2 variable so that I can graph this on my line graph?

[Nick Cox]

@Maarten Buis' point was that the regression calculates the means for you. That's what a regression is, or does, gives the mean outcome as a function of predictors. Do work through his worked example, which is deliberately one that you should be able to run yourself in your Stata.

A different point is that I would expect NO2 to be better analysed on logarithmic scale. The mean of log NO2 is equivalent to the geometric mean of NO2.