Factors
Using the 'Factors' step in the builder to create a table of risk factors, and calculate final ratings
Last updated
Using the 'Factors' step in the builder to create a table of risk factors, and calculate final ratings
Last updated
With the 'Factors' step, you can create a table of risk factors, and build calculations that will give you a final risk weightings or rules that you can use in pricing.
How might you use the 'Factor' step?
If you were building a pricing model for insurance, you'll have ingested a lot of data points by the time a customer has inputted their quote information and it's been enriched by other sources.
In order to calculate the final price, you'll need to simplify all of that data into some weightings or rules relating to the risk of the driver or the vehicle they're insuring.
For example, you will need to calculate a vehicle peril rating, taking into account the make of vehicle, its age, engine power and other factors.
Once you've added the 'Factor' step and named it, like you will with any other step in the Build tool (more on that here), you'll need to start adding any factor tables that you'll use to calculate the final risk factor in this step.
To get started, click the cog symbol in the top right of the 'Factor' step.
Once you've opened the step, scroll down to the 'Factor Table' and click the 'ADD' button in the top right.
Name the factor table you're adding. This will be the key that you use in all subsequent expressions and calculations, so it will need to be unique, descriptive, and easy to understand.
You can use letters, symbols and numbers, but no spaces.
Factor tables can include just one dimension (for example, vehicle engine capacity), or combine multiple related dimensions in the calculation's result (for example, vehicle year of manufacture against the number of year's No Claims Discount held by the driver).
The columns you can include in your factor table are:
Input - The key that you've set for your factor table.
Value - The value you're rating against. This could be standalone if it were a car make, for example (and you wouldn't need to fill in the 'Value to' column). However, if you wanted to return risk ratings based on numerical ranges, this value would need to be the bottom of the range.
Value to - This column is optional, but if you want to return results based on a numeric range, this will be the top value of your range.
Result - The risk rating that you want to return for the value on each row.
Excess - The excess that you want to return for the value on each row.
Exclude - The decision on whether you'll choose to accept or reject the value on each row. This will need to be set to FALSE if you want to accept the row, or TRUE if you'd like to exclude it.
You can then copy a data table and paste it directly into Swallow, or you can drag-and-drop a CSV document, which will upload immediately.
Once successfully uploaded, you'll be able to review the table to see that all the data has been uploaded correctly.
You can search for specific values in the table by clicking in the 'Filter Column' row beneath each column title and typing. The table will then filter automatically to show you results matching your search.
In the example below, the table shows the different rating factors of vehicle engine capacity.
To bulk edit the table, you can use the buttons to the top left of the table to download as a CSV, edit and then re-upload your factor table.
Once you're happy with your factor table. Click the 'Save & Close' button.
Once all of your factor tables are uploaded, you'll be left with a full table listing all the different factor tables you've added. It will list out the factor keys, the number of rows in each table, the default value (which is always set as 1), and the number of dimensions used in each table.
You can re-edit any table, or use the search bar in the top right to find a specific table if you've added many.
Once you've successfully uploaded all of your factor tables, you can write the final calculation for the step, completing all the fields below.
This is the name for your risk factor rule or weighting. This will be the key that you use in all subsequent expressions and calculations, so it will need to be unique, descriptive, and easy to understand.
You can use letters, symbols and numbers, but no spaces.
Write an expression to calculate your final risk factor, using any factors present in your model, including the factor tables you've uploaded.
Here's an example of how you would structure an expression to calculate a vehicle peril rating from vehicle make, vehicle year of manufacture, vehicle year of manufacture against No Claims Discount and vehicle engine capacity, rounded to 4 decimal places:
round({{vehicle_make_factor}} * {{vehicle_year_factor}} * {{vehicle_year_x_ncd_factor}} * {{vehicle_engine_capacity_factor}} , 4)
To learn how to structure your formula in more detail, debug and to see the full glossary, check out our full page on Syntax.
This will be the default value that your quote and tests will fall back on if for any reason the calculation doesn't function as expected and return a number. You may want to set this as a worst-case scenario.
Once you're happy with your factor calculation, just press the 'Save' button in the top right of the screen. You're now ready to head back to the 'Build' screen.
