How do I create a bar chart in python ggplot?

Question

I'm using yhat's ggplot library. I have the following pandas DataFrame:

   degree  observed  percent observed  expected  percent expected
0       0         0               0.0         0          0.044551
1       1         1               0.1         1          0.138604
2       2         3               0.3         2          0.215607
3       3         4               0.4         2          0.223592
4       4         1               0.1         2          0.173905
5       5         1               0.1         1          0.108208

At the moment, I'm doing the following (where df returned in the first line in the first line in the function is the DataFrame above):

def chartObservedExpected(graph):
    df = observedExpected(graph)
    return ggplot(aes(x='degree', y='percent observed'), data=df) + \
           geom_point() + \
           ylim(-0.015,1.015) + \
           xlim(-0.05,max(df['degree']) + 0.25) + \
           labs("Degree","Proportion of Total") + \
           ggtitle("Observed Node Degree Distribution")

chartObservedExpected(G)

This is what I get:

scatter

However, whenever I try geom_bar() instead of geom_point(), I end up with just 100% bars. I've tried just plain geom_bar() and also geom_bar(stat="percent observed"), but neither seem to work. This is always what I get:

bar

What I'm trying to do is to mimic/reproduce the following:

attempt

Any idea how to get the bar part working (or the whole thing, for that matter)?

Looks like a bug/something that hasn't been ported over yet. See the issue on Github [here](https://github.com/yhat/ggplot/issues/152). If it was working, it'd be something like `ggplot(aes(x='degree'), df) + geom_bar(aes(y='percent_observed'), stat='identity')`, but it looks like `stat`s other than `'bin'` don't work yet. — Marius, Mar 24 '14 at 01:45
@Marius Ok, thanks. Not that big of a deal for now, but I hope they implement it soon! — Clay, Mar 24 '14 at 02:10

score 24 · Accepted Answer · answered Mar 24 '14 at 03:05

24

Use weight, here is an example:

from ggplot import *
import pandas as pd
df = pd.DataFrame({"x":[1,2,3,4], "y":[1,3,4,2]})
ggplot(aes(x="x", weight="y"), df) + geom_bar()

the output looks like:

enter image description here

answered Mar 24 '14 at 03:05

HYRY

94,853
25
187
187

1

Thanks! Looks like this is a little funky if you try to use `xlim` to set the range of the x-axis, but it seems to work well otherwise. I also was unable to change the width of the bars. – Clay Mar 24 '14 at 14:54
1

Spent so much time trying to do this with `stat='identity'` and it didn't work. This does. – Jaycal Jan 29 '17 at 21:56

How do I create a bar chart in python ggplot?

1 Answers1

Linked