coefplot
Draws the coefficients of the model you just fitted, with their error bars.
What we have
- the coefficients of the last
regress,logit,probit,poissonornbreg - 95% intervals,
constant,keep()anddrop()
Coming later
- plotting two models side by side
- a
marginsplot, sincemarginsitself is not built
How you write it
coefplot [, constant] [keep(terms)] [drop(terms)]
Square brackets mean optional. A vertical bar means choose one.
What it does
A table of twenty coefficients is hard to read; the same twenty as points with 95% intervals, against a line at zero, takes a second. Anything whose interval crosses zero is not distinguishable from no effect.
It reads the last regress, logit, probit, poisson or nbreg, so run it straight afterwards. The constant is left out unless you ask, and keep() or drop() let you exclude coefficients measured on wildly different scales, which would otherwise squash the rest flat.
Try it
Every example below was run after sysuse vira.sample02, clear.
. regress mpce age education
Linear regression Number of obs = 6,000
F statistic = 277.09
Prob > F = 0.0000
R-squared = 0.0846
Root MSE = 1420.652
------------------------------------------------------------------------------
mpce | Coefficient Std. err. t P>|t| [95% conf. interval]
-------------+----------------------------------------------------------------
age | 6.137629 .9780934 6.28 0.000 4.220214 8.055044
education | 91.08834 3.980503 22.88 0.000 83.28513 98.89156
_cons | 2743.849 58.66231 46.77 0.000 2628.849 2858.848
------------------------------------------------------------------------------
The same thing in pandas and in R
pandas
import statsmodels.formula.api as smf
model = smf.ols("mpce ~ age + education", data=df).fit()
R
model <- lm(mpce ~ age + education, data = df) summary(model) # This R was run in a real R (4.6.1) and gave the same numbers.
. coefplot
(graph saved to graph_coefplot.html)
The same thing in pandas and in R
pandas
# rendered with D3; see graph_coefplot.html
R
# no R form for `coefplot` yet -- see docs/r-equivalents.md