Development lags could not be determined

[Coomaren]

Are you on the latest chainladder version?

• Yes, this bug occurs on the latest version.

Describe the bug in words

I already have my development in months. and I am trying to create a triangle as below. But I am getting an error.
It seems that this issue was discussed before. I apologise if I have missed this. I have just installed chainladder and so have the latest version

How can the bug be reproduced?

data = {
'origin': [2018, 2018, 2018, 2018, 2019, 2019, 2019, 2019],
'dev': [1, 2, 3, 4, 1, 2, 3, 4], # months
'value': [100, 80, 60, 40, 120, 90, 70, 50]
}
df = pd.DataFrame(data)
tri = cl.Triangle(df, origin='origin', development='dev', columns='value')

What is the expected behavior?

I would expect the triangle with the years plus the value at months 1-4

[henrydingliu]

The dataset as given is not a triangle, and not suitable for analysis.

Please refer to the tutorial for examples of chainladder-ready, triangle-shaped data.

[Coomaren]

It is true, I have not shared my data and used a dummy data.

My real data is industry data and was used for reserving in ResQ

The point is that my development period is in months whereas the Underwriting period is a year. So I am getting the same error

[henrydingliu]

Apologies for closing the issue inadvertently.

My real data is industry data and was used for reserving in ResQ

You can mask your original data by multiplying against a random vector.

The point is that my development period is in months whereas the Underwriting period is a year. So I am getting the same error

Please refer to the tutorial for examples of chainladder-ready data.

[kennethshsu]

Thanks @henrydingliu. And to add on more, @Coomaren you may want to try this:

import chainladder as cl
import pandas as pd

data = {
'origin': [2018, 2018, 2018, 2018, 2019, 2019, 2019, 2019],
'dev': [2018, 2019, 2020, 2021, 2019, 2020, 2021, 2022], # <-- these should be valuation dates
'value': [100, 80, 60, 40, 120, 90, 70, 50]
}

df = pd.DataFrame(data)
tri = cl.Triangle(df, origin='origin', development='dev', columns='value', cumulative = False)
tri

But yes, I would recommend following the tutorial @henrydingliu shared if you haven't.