Chapter 22 Appendix. Data Resources

This appendix section describes the datasets used in this book and others that you may wish to explore.

For each set of data, we provide download buttons so that you can easily access the data in standard .csv (comma separated value) format. This allows you replicate and experiment with the methods developed in the book as well as sharpen your understanding through exercises.

We provide the source of each dataset. We also recommend, for deeper understanding, that you occasionally refer to these original sources to further develop your appreciation of the data underpinning the analytics developed in this book.

22.1 Wisconsin Property Fund

Description: The Wisconsin Local Government Property Insurance Fund (LGPIF) is an insurance pool administered by the Wisconsin Office of the Insurance Commissioner. The LGPIF was established to provide property insurance for local government entities that include counties, cities, towns, villages, school districts, and library boards. The fund insures local government property such as government buildings, schools, libraries, and motor vehicles. It covers all property losses except those resulting from flood, earthquake, wear and tear, extremes in temperature, mold, war, nuclear reactions, and embezzlement or theft by an employee.

The data are available using this download button:

Table 22.1: Variables in the Wisconsin Property Fund Dataset

Variable	Description
PolicyNum	Policy number
Year	Contract year
Premium	Premium
Deduct	Deductible
BCcov	Coverage for building and contents
Freq	Number of claims during the year (frequency)
Fire5	Binary variable to indicate the fire class is below 5
NoClaimCredit	Binary variable to indicate no claims in the past two years
EntityType	Categorical variable that is one of six types: 1=Village, 2=City,3=County, 4=Misc, 5=School, or Town)
AlarmCredit	Categorical variable that is one of four types: (0, 5, 10, or 15) for automatic smoke alarms in main rooms
BCClaim	Builing and contents claims

Table 22.2: Wisconsin Property Fund First Five Rows

PolicyNum	Year	Premium	Deduct	BCcov	Freq	Fire5	NoClaimCredit	EntityType	AlarmCredit	BCClaim
120002	2006	9313	1000	22714456	0	1	0	3	1	0.00
120002	2007	8767	1000	25046646	0	1	0	3	1	0.00
120002	2008	7090	1000	20851525	0	1	1	3	1	0.00
120002	2009	8522	1000	21852696	0	1	1	3	1	0.00
120002	2010	7994	1000	23511493	1	1	1	3	1	6838.87

Table 22.2: Wisconsin Property Fund Last Five Rows

PolicyNum	Year	Premium	Deduct	BCcov	Freq	Fire5	NoClaimCredit	EntityType	AlarmCredit	BCClaim
180787	2010	199	500	285000	0	1	1	4	1	0.00
180788	2010	58344	100000	416739800	1	1	0	4	1	168304.05
180789	2010	295	500	500988	1	1	0	4	1	1034.33
180790	2010	2077	1000	3580665	0	1	0	4	4	0.00
180791	2010	81	500	118800	0	1	0	4	1	0.00

22.2 ANU Corporate Travel Data

Universities purchase corporate travel policies to cover employees and students traveling on official university business for a wide variety of accidents and incidents while away from the campus or primary workplace. This broad coverage includes medical care and evacuation, loss of personal property, extraction for political and weather related reasons, and more. See Frees and Butt (2022) for more information about this coverage.

There are 2107 observations in this dataset. The variable names are described in Table 22.3 and the first and last five observations are in Table 22.4.

Data are available using this button: .

Table 22.3: Variables in the Corporate Travel Dataset

Variable	Description
UW Year	Underwriting Year
Loss Date	Date that the loss occurred
Reported Date	Date that the loss was reported
Last Trans Date	Last date in which there was a transaction regarding the loss
Paid Loss	Cumulative amount paid on the loss
Outstanding Reserve	Estimate of the loss amount yet to be paid
Incurred Loss	Sum of the amount paid and the estimate of future payments
Status	An indicator as to whether the claim has been deemed settled (closed) or not settled (open)

Table 22.4: Corporate Travel Data First Five Rows

UW.Year	Loss.Date	Reported.Date	Last.Trans.Date	Paid.Loss	Outstanding.Reserve	Incurred.Loss	Status
2021	19/12/2021	20/12/2021	24/12/2021	10000.00	0	10000.00	Closed
2021	9/4/2022	29/04/2022	30/05/2022	423.08	0	423.08	Closed
2021	2/5/2022	4/5/2022		0.00	500	500.00	Open
2021	5/5/2022	17/05/2022		0.00	562	562.00	Open
2021	30/04/2022	27/05/2022	10/6/2022	1500.00	0	1500.00	Closed

Table 22.4: Corporate Travel Data Last Five Rows

UW.Year	Loss.Date	Reported.Date	Last.Trans.Date	Paid.Loss	Outstanding.Reserve	Incurred.Loss	Status
2006	1/11/2006	19/06/2007		0.00	0	0.00	Closed
2006	24/06/2007	26/06/2007	8/1/2008	6278.10	0	6278.10	Closed
2006	4/7/2007	6/7/2007	11/9/2007	114.50	0	114.50	Closed
2006	20/05/2007	26/06/2007	14/07/2007	135.65	0	135.65	Closed
2006	15/02/2007	27/06/2007	14/07/2007	1207.75	0	1207.75	Closed

Source: Frees, Edward and Butt, Adam (2022). “ANU Corporate Travel Insurance Claims 2022.” Australian National University Data Commons. DOI https://doi.org/10.25911/vrdw-9f32.

22.3 ANU Group Personal Accident Data

Group personal accident insurance offers financial protection in case of injury or death resulting from an incident that occurs on the job. Like workers’ compensation, group personal accident offers insurance coverage and liability insurance protection against accidental death or injury. Unlike workers’ compensation, group personal accident covers students and ANU’s voluntary workers. See Frees and Butt (2022) for more information about this coverage.

There are 148 observations in this dataset. The variable names are described in Table 22.5 and the first and last five observations are in Table 22.6.

Data are available using this button: .

Table 22.5: Variables in the Group Personal Accident Dataset

Variable	Description
UW Year	Underwriting Year
Loss Date	Date that the loss occurred
Last Trans Date	Last date in which there was a transaction regarding the loss.
Paid Loss	Cumulative amount paid on the loss
Outstanding Reserve	Estimate of the loss amount yet to be paid
Incurred Loss	Sum of the amount paid and the estimate of future payments
Status	An indicator as to whether the claim has been deemed settled (closed) or not settled (open)

Table 22.6: Group Personal Accident Data First Five Rows

UW.Year	Loss.Date	Last.Trans.Date	Paid.Loss	Outstanding.Reserve	Incurred.Loss	Status
2021	6/12/2021	3/6/2022	805.0	0.0	805	Closed
2021	15/11/2021		0.0	0.0	0	Closed
2021	15/11/2021		0.0	0.0	0	Closed
2021	22/03/2022	4/5/2022	396.0	0.0	396	Closed
2021	11/4/2022	2/8/2022	740.1	359.9	1100	Open

Table 22.6: Group Personal Accident Data Last Five Rows

UW.Year	Loss.Date	Last.Trans.Date	Paid.Loss	Outstanding.Reserve	Incurred.Loss	Status
2010	6/3/2011	26/07/2011	776.00	0	776.00	Closed
2010	22/07/2011	23/01/2012	4624.54	0	4624.54	Closed
2010	5/6/2011	30/01/2012	1503.65	0	1503.65	Closed
2007	11/1/2008	23/02/2008	0.00	0	0.00	Closed
2007	29/08/2008		0.00	0	0.00	Closed

Source: Frees, Edward and Butt, Adam (2022). “ANU Group Personal Accident Claims 2022.” Australian National University Data Commons. https://doi.org/10.25911/jcfx-zj56.

22.4 ANU Motor Vehicle Data

This policy covers ANU’s vehicles including cars, vans, utilities, and motorcycles. See Frees and Butt (2022) for more information about this coverage.

There are 318 observations in this dataset. The variable names are described in Table 22.7 and the first and last five observations are in Table 22.8.

Data are available using this button: .

Table 22.7: Variables in the Motor Vehicle Dataset

Variable	Description
Policy Term Start Date	Start date of the contract year in which the loss occurred
Loss Date	Date that the loss occurred
Reported Date	Date that the loss was reported
Motor Fault	Party responsible for the loss
Driver Age	Age of the driver
Vehicle Description	Type of vehicle
Loss Postcode	Postal code where the loss occurred
Excess	The deductible applied to the loss
Motor Net Paid	Amount paid to the insured (ANU)
Outstanding Estimate	Estimate of the loss amount yet to be paid
Motor Net Incurred	Sum of the amount paid and the estimate of future payments
Third Party Identified	Indicates whether a responsible third party could be identified
Third Party Insured	Indicates whether a responsible third party was insured

Table 22.8: Motor Vehicle Data First Five Rows

Policy.Term.Start.Date	Loss.Date	Reported.Date	Motor.Fault	Driver.Age	Vehicle.Description	Loss.Postcode
1/11/2011	6/6/2012	4/10/2012	THIRD PARTY RESPONSIBLE	NA	FORD TRANSIT VAN	2600
1/11/2011	16/08/2012	14/11/2013	INSURED RESPONSIBLE	39	TOYOTA HIACE	2612
1/11/2011	4/9/2012	17/01/2013	INSURED RESPONSIBLE	52	HYUNDAI IX35	2600
1/11/2011	21/09/2012	28/09/2012	THIRD PARTY RESPONSIBLE	59	HOLDEN COMMODORE	2518
1/11/2011	22/09/2012	12/10/2012	INSURED RESPONSIBLE	NA	SUBARU FORESTER	2612

Excess	Motor.Net.Paid	Outstanding.Estimate	Motor.Net.Incurred	Third.Party.Identified	Third.Party.Insured
1000	384.88	0	384.88	IDENTIFIED
1000	901.21	0	901.21
1000	1225.71	0	1225.71
NA	1671.76	0	1671.76	IDENTIFIED	NOT INSURED
1000	3418.86	0	3418.86		INSURED

Table 22.8: Motor Vehicle Data Last Five Rows

Policy.Term.Start.Date	Loss.Date	Reported.Date	Motor.Fault	Driver.Age	Vehicle.Description	Loss.Postcode
1/11/2021	4/4/2022	5/4/2022	INSURED RESPONSIBLE	66	VOLKSWAGEN TIGUAN	2604
11/1/2021	11/4/2022	9/5/2022	INSURED RESPONSIBLE	27	TOYOTA HILUX	2540
1/11/2021	11/4/2022	9/5/2022	INSURED RESPONSIBLE	27	TOYOTA HILUX	2540
11/1/2021	15/04/2022	11/7/2022	INSURED RESPONSIBLE	21	TOYOTA HILVX	2601
1/11/2021	18/07/2022	18/07/2022	NO-ONE RESPONSIBLE	NA	TOYOTA HILUX	2601

Excess	Motor.Net.Paid	Outstanding.Estimate	Motor.Net.Incurred	Third.Party.Identified	Third.Party.Insured
0	2373.49	1056.00	3429.49
0	210.00	25000.00	25210.00
0	0.00	31927.27	31927.27
0	0.00	2750.00	2750.00
0	0.00	299.00	299.00

Source: Frees, Edward and Butt, Adam (2022). “ANU Motor Vehicle Claims 2022.” Australian National University Data Commons. DOI https://doi.org/10.25911/g7e4-9e46.

22.5 Spanish Personal Insurance Data

This dataset consists of 10,000 insurance private customers of a real portfolio of insurance policy holders in Spain with a motor insurance and a homeowners insurance contract for policy year 2014. The data contain information on each customer, policies and yearly claims by type of contract.

The data are available using this download button:

The description of the data is the following:

Table 22.9: Variable and Description of Spanish Personal Insurance Data

Variable	Description
gender	1 for male and 0 for female
Age_client	the age of the customer in years
year	Policy year. Equals 5 corresponding to 2014.
age_of_car_M	the number of years since the vehicle was bought by the customer
Car_power_M	the power of the vehicle
Car_2ndDriver_M	1 if the customer has informed the insurance company that a second occasional driver uses the vehicle, and 0 otherwise
num_policiesC	the total number of policies held by the same customer in the insurance company
metro_code	1 for urban or metropolitan and 0 for rural
Policy_PaymentMethodA	1 for annual payment and 0 for monthly payment in the motor policy
Policy_PaymentMethodH	1 for annual payment and 0 for monthly payment in the homeowners policy
Insuredcapital_content_re	the value of content in homeowners insurance
Insuredcapital_continent_re	the value of building in homeowners insurance
appartment	1 if the homeowners insurance correspond to an apartment and 0 otherwise
Client_Seniority	the number of years that the customer has been in the company
Retention	1 if the policy is renewed and 0 otherwise
NClaims1	the number of claims in the motor insurance policy for the corresponding year
NClaims2	the number of claims in the homeowners insurance policy for the corresponding year
Claims1	the sum of claims cost in the motor insurance policy for the corresponding year
Claims2	the sum of claims cost in the homeowners insurance policy for the corresponding year
Types	1 when neither an auto nor a home claim, it is equal to 2 when the customer has an auto but not a home claim, it is equal to 3 when the customer does not have not an auto but a home claim and it is equal to 4 when both an auto and a home claim.
PolID	Policy Identification Number

All monetary units are expressed in Euros. In motor insurance, only claims at fault are considered.

Table 22.10: Spanish Personal Insurance Data First Five Rows

gender	Age_client	year	age_of_car_M	Car_power_M	Car_2ndDriver_M	num_policiesC	metro_code	Policy_PaymentMethodA	Policy_PaymentMethodH
1	47	5	12	163	0	0	0	1	1
1	52	5	13	80	0	1	0	1	1
0	66	5	7	97	0	1	1	1	1
1	70	5	17	95	0	1	0	1	1
1	67	5	13	110	0	1	0	1	1

Insuredcapital_content_re	Insuredcapital_continent_re	appartment	Client_Seniority	Retention	NClaims1	NClaims2	Claims1	Claims2	Types	PolID
10.18920207	12.0732153	1	6.58179329	1	0	0	0	0.00	1	12476
9.57144250	11.4431860	0	18.48049281	1	0	0	0	0.00	1	29232
9.33021633	11.2760512	1	15.08555784	1	0	0	0	0.00	1	23770
10.48489670	11.1366224	1	15.52361396	1	0	1	0	57.97	3	8228
10.96130962	12.3475909	0	6.10814511	1	0	0	0	0.00	1	37088

Table 22.10: Spanish Personal Insurance Data Last Five Rows

gender	Age_client	year	age_of_car_M	Car_power_M	Car_2ndDriver_M	num_policiesC	metro_code	Policy_PaymentMethodA	Policy_PaymentMethodH
1	66	5	8	143	0	1	0	1	1
1	55	5	18	125	1	1	0	1	1
0	41	5	10	190	0	1	0	1	1
1	50	5	5	140	0	1	0	1	1
1	55	5	12	90	0	1	1	1	1

Insuredcapital_content_re	Insuredcapital_continent_re	appartment	Client_Seniority	Retention	NClaims1	NClaims2	Claims1	Claims2	Types	PolID
10.30518223	11.4037721	1	19.73169062	1	0	0	0	0	1	2967
10.88842000	11.0707182	1	15.33470226	1	0	0	0	0	1	9387
9.22486636	11.6327216	1	6.00684463	1	0	0	0	0	1	36519
9.96916293	12.2717054	0	8.39151266	1	0	0	0	0	1	33276
11.12727769	12.7367038	0	6.42299795	1	0	0	0	0	1	25370

These data were drawn from a larger database of 40,284 insurance private customers. These customers are tracked from 2010 to 2014. Some customers do not renew their policies, so that they do not stay in the sample for five years. For the smaller data, only the 2014 policy year was used and from this, a random sample of 10,000 customers was drawn.

See Frees et al. (2021) for more information about this dataset. The larger database contains 122935 rows and is freely available at:

Source:

Guillen, Montserrat; Bolancé, Catalina; Frees, Edward W.; Valdez, Emiliano A. (2021), “Insurance data for homeowners and motor insurance customers monitored over five years,” Mendeley Data, V1, DOI https://doi.org/10.17632/vfchtm5y7j.1

22.6 ‘R’ Package CASdatasets

The R package CASdatasets provides a convenient way to access many well-known insurance datasets. This package was originally created to support the book Computational Actuarial Science with R, edited by Arthur Charpentier, Charpentier (2014).

To install the package, here is a bit of R code:

install.packages("CASdatasets", repos = "http://cas.uqam.ca/pub/", type = "source")
library(CASdatasets)
`?`(CASdatasets)
`?`(sgautonb  # See the documentation of the Singapore Auto Data
)
`?`(lossalae  # See the documentation of the Loss and Expense Data
)

Note that this package assumes that you have already installed a few other packages, including xts, sp, and zoo.

To illustrate,

• in Chapter 3 we use the Singapore data (referred to as sgautonb in the package) and
• in Chapter 16 we use the loss and expense data (referred to as lossalae in the package).

22.7 Other Data Sources

There exists man other (non-actarial) data sources. First, data can be obtained from university-based researchers who collect primary data. Second, data can be obtained from organizations that are set up for the purpose of releasing secondary data for the general research community. Third, data can be obtained from national and regional statistical institutes that collect data. Finally, companies have corporate data that can be obtained for research purposes.

While it might be difficult to obtain data to address a specific research problem or answer a business question, it is relatively easy to obtain data to test a model or an algorithm for data analysis. In the modern era, readers can obtain datasets from the Internet. The following is a list of some websites to obtain real-world data:

• UCI Machine Learning Repository. This website (url: http://archive.ics.uci.edu/ml/index.php) maintains more than 400 datasets that can be used to test machine learning algorithms.
• Kaggle. The Kaggle website (url: https://www.kaggle.com/) include real-world datasets used for data science competitions. Readers can download data from Kaggle by registering an account.
• DrivenData. DrivenData aims at bringing cutting-edge practices in data science to solve some of the world’s biggest social challenges. In its website (url: https://www.drivendata.org/), readers can participate in data science competitions and download datasets.
• Analytics Vidhya. This website (url: https://datahack.analyticsvidhya.com/contest/all/) allows you to participate and download datasets from practice problems and hackathon problems.
• KDD Cup. KDD Cup is the annual Data Mining and Knowledge Discovery competition organized by the ACM Special Interest Group on Knowledge Discovery and Data Mining. This website (url: http://www.kdd.org/kdd-cup) contains the datasets used in past KDD Cup competitions since 1997.
• U.S. Government’s open data. This website (url: https://www.data.gov/) contains about 200,000 datasets covering a wide range of areas including climate, education, energy, and finance.
• AWS Public Datasets. In this website (url: https://aws.amazon.com/datasets/), Amazon provides a centralized repository of public datasets, including some huge datasets.

Bibliography

Charpentier, Arthur. 2014. Computational Actuarial Science with r. CRC press.

Frees, Edward W, Catalina Bolancé, Montserrat Guillen, and Emiliano A Valdez. 2021. “Dependence Modeling of Multivariate Longitudinal Hybrid Insurance Data with Dropout.” Expert Systems with Applications 185: 115552.

Frees, Edward W, and Adam Butt. 2022. “ANU Insurable Risks.” https://doi.org/10.25911/0SE7-N746.