Why you should attend:
- You will build personal auto pricing models
- You’ll create visualizations of data including maps
- Hear from industry experts
- Learn using actual depersonalized data
Who should attend: This course is open to all. However, it is tailored to insurance modeling actuaries and will use samples of actual depersonalized insurance data.
Cost: FREE
Duration: We anticipate each lesson to occur monthly and be around 1 hour in length starting at 2:00 PM EST. Please see the syllabus for specific dates.
RESERVE YOUR SPOT
Lesson 1: Python vs other languages - a gentle introduction
You can view the recorded lesson here.
Topics Covered:
- List comprehension including enumeration
- Create dictionary and convert to dataframe
- Adjust dtype in the dataframe + basic function
- Basic dataframe statistics using .agg
Lesson 2: Working with basic data and visualization
You can view the recorded lesson here.
Topics Covered:
- Import data from .csv + cover other options (sql, pickle, feather)
- Rename columns + downcast to minimum size date type
- Display data using matplotlib + function
- Merge other data sources to dataset
Questions?
Please feel free to send any questions by email to [email protected].
Lesson 3: Exploratory data analysis and more visualization
You can view the recorded lesson here.
Topics Covered:
- Build EDA tool using ipywidgets
- Perform data cleanup & feature engineering using iterative function application and EDA process
Questions?
Please feel free to send any questions by email to [email protected].
Lesson 4: Creating maps with Python
You can view the recorded lesson here.
Topics Covered:
- Visualize geographic data using plotly and folium
- Build function & widget for EDA with geographic focus
Questions?
Please feel free to send any questions by email to [email protected].
Lesson 5: Advanced maps with Python
You can view the recorded lesson here.
Topics Covered:
- Edit geojson dictionary for better folium map experience
Questions?
Please feel free to send any questions by email to [email protected].
Lesson 6: Generalized linear models
You can view the recorded lesson here.
Topics Covered:
- Fit GLM with statsmodels
- Review holdout lift chart
- Build and analyze residual plots
- Review map of model results
- Double lift chart
Questions?
Please feel free to send any questions by email to [email protected].
Lesson 7: XGBoost gradient boosting machine models
You can view the recorded lesson here.
Topics Covered:
- Fit GBM with xgboost
- Run SHAP values and review model impacts
- Review holdout lift chart, residual plots, and map of model results
Questions?
Please feel free to send any questions by email to [email protected].
Lesson 8: Using Ray and Modin for faster processing
You can view the recorded lesson here.
Topics Covered:
- Modin and Ray packages speed up processing with parallelization
- Import large csv files quickly
- Parallelize data frame calculations
- Run multiple model grid search processes at once
Questions?
Please feel free to send any questions by email to [email protected].