catplot doesn´t display bars while importing a csv

Question

I have made my own code trying to find any differences but I don´t know where is the trick, here is one catplot that display well:

import matplotlib.pyplot as plt
import seaborn as sns
import pandas as pd

# Change the color palette to "RdBu"
par_ad = {
    "Parents advice":[1,2,3,3,4,5,4,3,1,1,1,1,1,2,3,5,2,1,3,4]
}
df = pd.DataFrame(par_ad)
sns.set_style("whitegrid")
sns.set_palette("RdBu")


# Create a count plot of survey responses
sns.catplot(x="Parents advice", 
            data=df, 
            kind="count")

# Show plot
plt.show()

On the other hand I have been working with a .csv and using orders inside it. Why it doesn´t show the bars? The code, the .csv belongs to a Datacamp course about Seaborn, it doesn´t throw any error apart from not displaying the bars. First line in .csv is the categories, where one of them is "Parents' advice" and the other lines are ints between 1 to 5 that are equivalent to the number of times where students ask for help from their parents.

import seaborn as sns
import pandas as pd

# Change the color palette to "RdBu"
survey_data = pd.read_csv("young-people-survey-responses.csv")
sns.set_style("whitegrid")
sns.set_palette("RdBu")

# Create a count plot of survey responses
category_order = ["Never", "Rarely", "Sometimes", 
                  "Often", "Always"]

sns.catplot(x="Parents' advice", 
            data=survey_data, 
            kind="count", 
            order=category_order)

# Show plot
plt.show()

Answer 1

How is sns supposed to know that "Never" equals 1, "Rarely" equals 2, and so on?

Option 1

• Pass range(1,6) to the order parameter and use ax.set_xticklabels to overwrite the range values.

import seaborn as sns
import matplotlib.pyplot as plt
import pandas as pd

# using the values from your initial example
par_ad = {"Parents' advice": [1,2,3,3,4,5,4,3,1,1,1,1,1,2,3,5,2,1,3,4]}
survey_data = pd.DataFrame(par_ad)

sns.set_style("whitegrid")
sns.set_palette("RdBu")

category_order = ["Never", "Rarely", "Sometimes", 
                  "Often", "Always"]

ax = sns.catplot(x="Parents' advice", 
                 data=survey_data, 
                 kind="count", 
                 order=range(1,6)) # set order: 1, 2, 3, 4, 5

# adjust the labels
ax.set_xticklabels(category_order)

plt.show()

Result:

Option 2

• Use Series.map to change the range values into your string values and make the column categorical using pd.Categorical before passing it to sns.catplot.

# ...

category_order = ["Never", "Rarely", "Sometimes", 
                  "Often", "Always"]

# create a dict to map cats to unique values in the `"Parents' acvice"` column, dropping any nan values
unique_advice = sorted(survey_data["Parents' advice"].dropna().unique())
mapper = dict(zip(unique_advice, category_order))

# make column categorical 
survey_data["Parents' advice"] = (pd.Categorical(survey_data["Parents' advice"]
                                                 .map(mapper), 
                                                 categories=category_order))

ax = sns.catplot(x="Parents' advice", 
                 data=survey_data, 
                 kind="count") # `order` no longer needed

plt.show()
# same result