Time series forecast with customer numbers

Hello there,

since I am new to knime I need some help.
I want to create a forecast with the following information: Date and customer numbers: Each customer number has its own time series with additional information such as new loans sold, bank account balances and account transactions per month, prime rate and inflation. The dataset also includes customers who have not (yet) received a loan.
However, the sold credits are very rare and therefore give some zero values.

Therefore the question if someone can explain me how to forecast the sold credits for the next 12 months. Is it necessary to split the data set so that I get one data set or time series per customer number? I would like to avoid this, as the number of customer numbers in my dataset is quite high.

Unfortunately, I have not yet been able to find a solution that works for me. I hope someone can help me here.
Thanks a lot!

@Molly123 welcome to the KNIME forum. Time Series are sometimes a challenge. One starting point could be this book by KNIME about the very topic.

Codeless Time Series Analysis with KNIME
A practical guide to implementing forecasting models for time series analysis applications

Then the Python package prophet is often used with time series. I would have to check if I find a suitable example that would cover the additional data provided. One thing you could also do is check sites like Kaggle to see if there is a similar task with publicly available data that might server as an example that also can be discussed without spelling any secrets.

1 Like

This topic was automatically closed 7 days after the last reply. New replies are no longer allowed.