How do I plot a histogram of a non-numerical column w.r.t another calculated column in pandas dataframe?

Question

I have a dataframe:

    index   zip lat lng city    state_id    state_name  population  density county_name timezone
0   0   01001   42.0626 -72.6259    Agawam  MA  Massachusetts   16769   565.8   Hampden America/New_York
1   1   01002   42.3749 -72.4621    Amherst MA  Massachusetts   29049   203.8   Hampshire   America/New_York
2   2   01003   42.3919 -72.5248    Amherst MA  Massachusetts   10372   5629.7  Hampshire   America/New_York
3   3   01005   42.4202 -72.1061    Barre   MA  Massachusetts   5079    44.3    Worcester   America/New_York
4   4   01007   42.2787 -72.4003    Belchertown MA  Massachusetts   14649   107.4   Hampshire   America/New_York
... ... ... ... ... ... ... ... ... ... ... ...
460 531 02771   41.8379 -71.3174    Seekonk MA  Massachusetts   13708   288.1   Bristol America/New_York
461 532 02777   41.7570 -71.2121    Swansea MA  Massachusetts   15840   269.7   Bristol America/New_York
462 533 02779   41.8349 -71.0754    Berkley MA  Massachusetts   6411    149.9   Bristol America/New_York
463 534 02780   41.9076 -71.1196    Taunton MA  Massachusetts   49036   573.1   Bristol America/New_York
464 535 02790   41.5999 -71.0832    Westport    MA  Massachusetts   15717   113.0   Bristol America/New_York
465 rows × 11 columns

I have to plot a histogram of cities and their populations. So, I used the following code from this answer:

import pylab as plt
ma_hist = ma_StateData.hist('city',weights=ma_StateData['population'] )
plt.ylabel('population')
plt.show()

This yields an error:

ValueError: hist method requires numerical columns, nothing to plot.

I also tried by the documentation:

ma_StateData.columns
# ma_histogram = pd.DataFrame.hist(ma_StateData, column='city', by='population')
# ma_histogram.plot.hist()
ma_city_population = ma_StateData[['city','population']]
ma_city_population.plot.hist(by='city')

But the resulting histogram is not correct(not what I am looking for):

What I want is a histogram whose x axis is cities, and corresponding to each city, the bar shows the population of that city.

How do I do that?

Answer 1

I only copied the first 5 rows of your data, I am not sure if you really need histogram, you could do it with groupby and sort_values:

>>> df.groupby('city')['population'].sum().sort_values(ascending=False).plot(kind='bar')