How can I dynamically create bins in Python?

Question

I have the following np array:

[['ID1', 922.63, 'Product 1'],
['ID1', 1001, 'Product 2'],
['ID1', 800, 'Product 1'],
['ID1', 922.63, 'Product 1'],
['ID1', 1001, 'Product 2'],
['ID2', 800, 'Product 1'],
['ID2', 922.63, 'Product 1'],
['ID2', 1001, 'Product 2'],
['ID3', 800, 'Product 1'],
['ID3', 700.63, 'Product 1'],
['ID3', 1200, 'Product 2'],
['ID3', 850, 'Product 1']]

The '2nd column' ($ amount) is what I care about. I want to build a histogram of product 1 and product 2, but I want the bins to be sized by 100. The actual data set I'm using has 75K rows and values that range from $1 to $200000. I want to automatically create these 'buckets' for the values and then build a histogram.

I thought it would be easy to find info on this using either pandas or numpy but I am either a newb and not able to understand other 'similar' solutions, or am just not finding what I'm looking for. Seems like it should be straight forward.

Each row is that one string or three seperate columns or values with a space as a delimeter? — Scott Boston, Nov 02 '17 at 21:09
I think [this question](https://stackoverflow.com/questions/47079717/pandas-group-by-equal-range) is very similar to that what you are looking for... — MaxU - stand with Ukraine, Nov 02 '17 at 21:11
@ScottBoston I'm really not sure. It starts as a dataframe and I convert it as such 'dataset = dataset.values'; print(dataset) — user3486773, Nov 02 '17 at 21:13

Sebastian Wozny · Accepted Answer · 2017-11-02T21:21:25.820

4

You can get a histogram by turning your data into a pandas.DataFrame:

a = [['ID1', 922.63, 'Product 1'],
['ID1', 1001, 'Product 2'],
['ID1', 800, 'Product 1'],
['ID1', 922.63, 'Product 1'],
['ID1', 1001, 'Product 2'],
['ID2', 800, 'Product 1'],
['ID2', 922.63, 'Product 1'],
['ID2', 1001, 'Product 2'],
['ID3', 800, 'Product 1'],
['ID3', 700.63, 'Product 1'],
['ID3', 1200, 'Product 2'],
['ID3', 850, 'Product 1']]
q=pd.DataFrame(a,columns=['id','price','product'])
q.hist(column='price',bins=100)

You can specify the number of bins you want with the bins parameter:

 q.hist(column='price', bins=100)

If you want to group it by product use the by parameter:

 q.hist(column='price', bins=100,by='product')

edited Nov 02 '17 at 21:21

answered Nov 02 '17 at 21:12

Sebastian Wozny

16,943
7
52
69

what does column=1 mean? Is that referencing the index of 1? – user3486773 Nov 02 '17 at 21:15
I modified the code to give the price column a real name – Sebastian Wozny Nov 02 '17 at 21:17
Awesome, one more thing, how could you separate this out and have two histograms based on each product? – user3486773 Nov 02 '17 at 21:19

How can I dynamically create bins in Python?

1 Answers1