Tools for processing and reporting on ILEC data, especially those used for
- Data preparation: e.g. to strip leading spaces from product names, save as faster-loading binary file in parquet format
- Preprocessing dataset published in 2021
- Preprocessing dataset published in 2017
- 2019 ValAct presentation on analytics efforts
- See sample notebook 2019_ValAct_demo.ipynb
- Beyond Actual to Table paper
- Related: 2017 Annual Meeting slides
- See sample notebook Beyond_Actual_to_Table.ipynb
- To do: 2017 Individual Life Insurance Mortality Experience Report
- duplicate figures... a notebook for this would be nice
Subpackages
data_prep: downloading, preprocessing 2021 datarollforward: rollforward computation and presentation, used for the 2019 ValAct presentationdecomp_trend: Trend of study year exposure or other metrics by decomposition, used for the 2019 ValAct presentationa2t
See the notebooks in the directory sample_notebooks for examples of use.
Please use an environment manager. To leave your environment untouched you can add the ILECTools directory to your system path:
import sys
sys.path.append('ILECTools') # specify the path as needed
import ilectools
# now you're good to go
An "environment" is a set of various packages, all of which have their own versions. Not all
package versions are compatible with each other so it is important to keep track of what
you are doing. The notebooks have been tested in an environment which conda-users
should be able to duplicate with the included environment specification files in the
env subdirectory. I intend for you to be able to create the ilec0 environment
on your machine with this command:
conda env create -f environment.yml # of course include any path you need to the yml file
These sample notebooks were run on machines with 32gb RAM, Ubuntu 20.04, with a 32gb swapfile. Your mileage may vary! You can see in the environment specs that Python 3.10.4 was used. Miniforge was used for environment management, as opposed to miniconda or the Anaconda python distribution. Pay attention to your licenses always.
For more on environment management with conda: