如何使用 displot 在 python 中制作 seaborn plot ，其中我们计算一个字段中的唯一值而不是总行数？

Question

I have a dataframe that contains about 60,000 rows. All 60,000 of them have unique record identifiers, but they also have separate sessionIDs, of which about 12,000 are unique.

I am trying to use seaborn distplot in order to make figures using these values, but when distplot does the aggregation, I can only get it to count the number of records and I cannot get it to aggregate over the number of unique sessionIDs.

Here is an example dataframe.

temp_df = pd.DataFrame([['d7d1b050-0e48-4c00-8061-c78817155b72',
  '42773088-e38f-4578-bc2a-69d1797a90eb',
  11,
  'groupA'],
 ['962c397d-a8f8-4f1c-a589-ecf74a7da62d',
  'b5baafb0-f6d4-4b4e-bc76-1287614b985d',
  10,
  'groupA'],
 ['a90fde40-9b9f-466e-bd5e-a40325b5fc9d',
  'b3fba007-aef5-4a5f-a53b-94eb0705d953',
  11,
  'groupB'],
 ['22ebb056-603c-4f66-8240-8c54e8043509',
  'b780fa66-addd-48c0-8db4-d755ebd351b8',
  10,
  'groupC'],
 ['52ffd64c-a5c1-4cd5-89c8-c1dcb8bd24b2',
  '37482cb7-c354-4b4b-92b6-2aaa62811e5b',
  10,
  'groupA'],
 ['55524169-f159-4c31-b939-bb00e1cba804',
  '34a9ff63-ea75-473d-ab89-9a92c3f4a8d9',
  10,
  'groupB'],
 ['2027d9d0-1e29-4d1f-969a-995a47f12052',
  '875488ea-85a2-47cb-b1ea-62003bbce80a',
  10,
  'groupA'],
 ['10d9c9fb-b5dd-4581-b148-a6198abecec1',
  '3f4b0604-513a-424b-98a3-e788ab3daa97',
  11,
  'groupD'],
 ['1c1e183b-6459-41bd-99aa-5f89b375006a',
  '53dd2ffd-c9b0-49c3-9275-190716c78799',
  10,
  'groupB'],
 ['31030ded-64a7-4854-8042-585605141e71',
  'f0514527-2d7b-4cad-a36f-f21e3425093c',
  10,
  'groupD'],
 ['cdfd5a0c-dd8c-4546-ba31-c2f021fb4859',
  '1ed007fe-d4f7-41bc-8f3c-b163c57f8a1f',
  11,
  'groupE'],
 ['66bd16a5-b514-4d8a-ad7a-afb8921f7dd2',
  'a2e9f137-bba5-46ec-8b13-7b17821de735',
  10,
  'groupB'],
 ['3cdb21d9-be3c-4723-bf28-0a7769d492b4',
  '9a6f1516-54a0-4dda-83d7-e05311e87ff5',
  10,
  'groupE'],
 ['d25f4cb2-3bf7-4898-a8a3-91d9e1b58576',
  '716a7732-6bcd-478d-87f9-c13cd83eaf66',
  11,
  'groupA'],
 ['e95134fd-7ce2-4e88-808c-e5abf13a4892',
  'c021c21b-7bab-4e1f-9ff0-4dfc584263b8',
  11,
  'groupE'],
 ['e13da005-1033-466f-b984-48fdfa0988f2',
  '5bcc0651-0775-4fa5-b521-ac90e0a33b1c',
  10,
  'groupB'],
 ['b60ee53d-e4fc-4e37-aa1c-df67f66e304e',
  '592adca4-6fa6-48c3-be97-2357250d736d',
  10,
  'groupD'],
 ['c1d47246-838f-418a-a92d-7b5150122775',
  'ff5d180c-cca9-474a-974e-e18c35cab912',
  10,
  'groupA'],
 ['fc129686-f7cd-407a-aca3-68f86c52af41',
  'a18dfc3a-2ce6-43f7-a21f-4c7371cff2b6',
  11,
  'groupE'],
 ['191af645-cb9e-408a-af2e-b6826f7177b9',
  'd430610b-b7da-42cb-aa93-c7f94774093c',
  10,
  'groupA']])

temp_df.columns = ['clickId', 'sessionId', 'month','group']

sns.displot(data=temp_df, x='month', hue='group')

Conceptually, I guess what I want to do is take the dataframe and eliminate all duplicate rows at the sessionId level, but I don't know how to do that.

Can someone help me?

Thanks, Brad

Answer 1

The answer is surprisingly simple.

When I was trying to draw the original plot, I was doing

sns.displot(temp_df, x='month', hue='group') which then included all of the data, so it was using unique row identifiers, but since I wanted to go with just sessionId, the solution I found was

sns.displot(temp_df[['sessionId', 'month','group']].drop_duplicates(), x='month', hue='group')

and that works.

Hopefully this helps someone else.