-
|
Hello, |
Beta Was this translation helpful? Give feedback.
Answered by
jbogaardt
Oct 13, 2021
Replies: 1 comment 1 reply
-
|
Hi @xne0n , Thanks for the question. It's not documented. We should probably mention something in the docs or figure out how to support wide. I'd use import chainladder as cl
# Suppose you have a wide triangle (perhaps copied from Excel clipboard...)
df = cl.load_sample('raa').to_frame()
# You can use pandas unstack to make wide format long
df = df.unstack().dropna().reset_index()
# make sure you have good column labels
df.columns = ['age', 'origin', 'values']
# and make a date-like valuation
df['valuation'] = (df['origin'].dt.year + df['age'] / 12 - 1).astype(int)
# Finally make a Triangle
cl.Triangle(df, origin='origin', development='valuation', columns='values') |
Beta Was this translation helpful? Give feedback.
1 reply
Answer selected by
xne0n
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Hi @xne0n ,
Thanks for the question. It's not documented. We should probably mention something in the docs or figure out how to support wide. I'd use
pandas.unstackto get there. Something like this: