CODEWITHSUNDEEP
pandasmatplotlibseaborn
Py.rookie89
Py.rookie89

Reputation: 129

Visualize countplot like linegraph python

I have a data frame for customer transactions:

     customer_id|tier |transaction_type|year|no_of_purchases
       1353455  |1    |online          |2012|5
       1353455  |1    |retail          |2012|8
       1353455  |1    |retail          |2014|10
       1543798  |2    |retail          |2012|1

The tier is their loyalty program tier. I want to see if there is difference in the transaction_type and no_of_purchases between each tier.

I did a countplot for this:

sns.countplot(x="transaction_type", y="no_of_purchase", hue"tier")

Goal: I want show for each year (x-axis) what is the count of the no_purchases for each transaction_type and how does this differ for each tier. Here is an example of how I want it to look. Ideally we would have 4 graphs, one for each of the transaction type. It would be great if we can show the count at each point too.

enter image description here

Upvotes: 1

Views: 7306

Answers (1)

dzakyputra
dzakyputra

Reputation: 682

You can use seaborn catplot.

First, you can aggregate your data, then do the vizualization using seaborn.

data_viz = df.groupby(['year','transaction_type','tier'], as_index=False)['no_of_purchases'].sum()

sns.catplot(data=data_viz, x='year', y='no_of_purchases', hue='tier', col='transaction_type', kind='bar')

But unfortunately based on the documentation, you cannot plot the data using line plot.

There is a work around though, you can do something like this.

import matplotlib.pyplot as plt

data_viz = df.groupby(['year','transaction_type','tier'], as_index=False)['no_of_purchases'].sum()

for i in list(data_viz['transaction_type'].unique()):
    viz = sns.lineplot(data=data_viz[data_viz['transaction_type'] == i], x='year', y='no_of_purchases', hue='tier')
    plt.title(i)
    plt.show()

Upvotes: 2

Related Questions