Bon Air Crime Analysis for February 2020. Reports are Slightly Down.

This is an analysis for February, 2020 based on data posted by the Louisville Metro Police Department at https://data.louisvilleky.gov/dataset/crime-reports/resource/6d81fdf8-27e7-41ac-9ae4-1bfd8730f002 

It is a Microsoft Excel CSV file that requires some manipulation and sorting and beginning this year it was loaded into JASP, which is an open source statistical package.  This analysis does include events that happen on the edge of the neighborhood.



As usual it includes the crimes reported within the boundaries of the historical Bon Air Neighborhood as defined:
1) Bardstown Road on the West  
2) Hikes Lane on the South
3) Furman Blvd on the East and
4) Taylorsville Rd and the Watterson Expressway on the North. 

The Analysis reflects events that happen on the lines and so some of the reports reflect what has happened just on the line and outside the geopolitical boundary of the neighborhood.   

The missing person reports at the 2300 block of Goldsmith Lane were removed. Otherwise, there were 46 reports within the boundaries.   There were 28 reports in the 40218 portion and 18 in the 40220 area. 

Variable		Level		Counts		Total		Proportion
ZIP_CODE		40218		28		46		0.609
		40220		18		46		0.391
Note.  Proportions tested against value: 0.5.

The following table shows the days of the week and what types of crimes were reported:

DAYS OF THE WEEK AND TYPES OF CRIME

Binomial Test
Variable		Level		Counts		Total		Proportion		p
WEEKDAY		FRIDAY		6		46		0.130		< .001
		MONDAY		7		46		0.152		< .001
		SATURDAY		7		46		0.152		< .001
		SUNDAY		3		46		0.065		< .001
		THURSDAY		8		46		0.174		< .001
		TUESDAY		9		46		0.196		< .001
		WEDNESDAY		6		46		0.130		< .001
CRIME_TYPE		ASSAULT		12		46		0.261		0.002
		BURGLARY		2		46		0.043		< .001
		DRUGS/ALCOHOL VIOLATIONS		2		46		0.043		< .001
		FRAUD		3		46		0.065		< .001
		MOTOR VEHICLE THEFT		4		46		0.087		< .001
		OTHER		9		46		0.196		< .001
		THEFT/LARCENY		7		46		0.152		< .001
		VANDALISM		2		46		0.043		< .001
		VEHICLE BREAK-IN/THEFT		5		46		0.109		< .001
Note.  Proportions tested against value: 0.5.

Tuesday was the day when the most police calls were made.  The most frequent type crime was assault.  The "other" reports included one domestic-violence report, but were otherwise the serving of a MIW and two missing person reports. 

The most common place for crime/police reports were residents. 

LOCATIONS OF POLICE REPORTS

Binomial Test
Variable		Level		Counts		Total		Proportion		p
PREMISE_TYPE		COMMERCIAL / OFFICE BUILDING		1		46		0.022		< .001
		CONVENIENCE STORE		2		46		0.043		< .001
		CYBERSPACE		1		46		0.022		< .001
		DEPARTMENT / DISCOUNT STORE		1		46		0.022		< .001
		GROCERY / SUPERMARKET		1		46		0.022		< .001
		HIGHWAY / ROAD / ALLEY		5		46		0.109		< .001
		HOTEL / MOTEL / ETC.		2		46		0.043		< .001
		OTHER / UNKNOWN		1		46		0.022		< .001
		OTHER RESIDENCE (APARTMENT/CONDO)		3		46		0.065		< .001
		PARKING LOT / GARAGE		3		46		0.065		< .001
		RESIDENCE / HOME		21		46		0.457		0.659
		RESTAURANT		1		46		0.022		< .001
		SCHOOL - ELEMENTARY / SECONDARY		1		46		0.022		< .001
		SERVICE / GAS STATION		1		46		0.022		< .001
		SPECIALTY STORE (TV, FUR, ETC)		2		46		0.043		< .001
Note.  Proportions tested against value: 0.5.

The block addresses of the crime reports are in the following table.  The 2700 block of Hikes Lane and the 3600 block of Stanton Blvd had the most reports with three each. 

Block Addresses of Crimes in Residences February 2020

Binomial Test
Variable		Level		Counts		Total		Proportion		p
BLOCK_ADDRESS		2200 BLOCK BASHFORD MANOR LN		1		21		0.048		< .001
		2300 BLOCK TERRIER CT		2		21		0.095		< .001
		2700 BLOCK HIKES LN		3		21		0.143		0.001
		2700 BLOCK MAXON DR		1		21		0.048		< .001
		2800 BLOCK MEADOW DR		2		21		0.095		< .001
		2800 BLOCK ROSALEE AVE		1		21		0.048		< .001
		3000 BLOCK TALISMAN RD		1		21		0.048		< .001
		3200 BLOCK ALLISON WAY		1		21		0.048		< .001
		3200 BLOCK WELLINGMOOR AVE		1		21		0.048		< .001
		3300 BLOCK BARDSTOWN RD		1		21		0.048		< .001
		3300 BLOCK TERRIER LN		1		21		0.048		< .001
		3500 BLOCK GOLDSMITH LN		1		21		0.048		< .001
		3500 BLOCK RAMONA AVE		1		21		0.048		< .001
		3600 BLOCK GREEN MEADOWS DR		1		21		0.048		< .001
		3600 BLOCK STANTON BLVD		3		21		0.143		0.001
Note.  Proportions tested against value: 0.5.

The reports at 3600 Stanton Blvd were all on one day and were domestic disturbance/violence issues. Also, the three reports at 2700 Hikes lane were simultaneously made at one time and appear to be part of one event at a residence.  

Assaults were the most common type of police reports made at residences.

Crime types in Bon Air Residences February 2020

Binomial Test
Variable		Level		Counts		Total		Proportion		p
CRIME_TYPE		ASSAULT		8		21		0.381		0.383
		BURGLARY		2		21		0.095		< .001
		MOTOR VEHICLE THEFT		1		21		0.048		< .001
		OTHER		5		21		0.238		0.027
		THEFT/LARCENY		2		21		0.095		< .001
		VANDALISM		1		21		0.048		< .001
		VEHICLE BREAK-IN/THEFT		2		21		0.095		< .001
Note.  Proportions tested against value: 0.5.

The assaults happened at the following locations of residences. 

Block Addresses of the Assaults in Bon Air Residences

Binomial Test
Variable		Level		Counts		Total		Proportion		p
BLOCK_ADDRESS		2300 BLOCK TERRIER CT		1		8		0.125		0.070
		2700 BLOCK HIKES LN		2		8		0.250		0.289
		2800 BLOCK MEADOW DR		2		8		0.250		0.289
		3300 BLOCK BARDSTOWN RD		1		8		0.125		0.070
		3600 BLOCK STANTON BLVD		2		8		0.250		0.289
Note.  Proportions tested against value: 0.5.

The 3300 block of Bardstown Road continues to be the location of the most police reports in the neighborhood with 10 reports. 

POLICE REPORTS BY ADDRESS

Binomial Test
Variable		Level		Counts		Total		Proportion		p
BLOCK_ADDRESS		2200 BLOCK BASHFORD MANOR LN		1		46		0.022		< .001
		2200 BLOCK GERALD CT		1		46		0.022		< .001
		2200 BLOCK GOLDSMITH LN		1		46		0.022		< .001
		2200 BLOCK HEATHER LN		1		46		0.022		< .001
		2200 BLOCK HIKES LN		1		46		0.022		< .001
		2300 BLOCK TERRIER CT		3		46		0.065		< .001
		2700 BLOCK HIKES LN		3		46		0.065		< .001
		2700 BLOCK MAXON DR		1		46		0.022		< .001
		2800 BLOCK DEL RIO PL		2		46		0.043		< .001
		2800 BLOCK MEADOW DR		3		46		0.065		< .001
		2800 BLOCK ROSALEE AVE		1		46		0.022		< .001
		2900 BLOCK HIKES LN		1		46		0.022		< .001
		3000 BLOCK TALISMAN RD		1		46		0.022		< .001
		3200 BLOCK ALLISON WAY		1		46		0.022		< .001
		3200 BLOCK WELLINGMOOR AVE		1		46		0.022		< .001
		3300 BLOCK BARDSTOWN RD		10		46		0.217		< .001
		3300 BLOCK FURMAN BLVD		1		46		0.022		< .001
		3300 BLOCK TERRIER LN		1		46		0.022		< .001
		3400 BLOCK BARDSTOWN RD		1		46		0.022		< .001
		3500 BLOCK BROCKTON LN		1		46		0.022		< .001
		3500 BLOCK GOLDSMITH LN		2		46		0.043		< .001
		3500 BLOCK RAMONA AVE		1		46		0.022		< .001
		3600 BLOCK GREEN MEADOWS DR		1		46		0.022		< .001
		3600 BLOCK STANTON BLVD		3		46		0.065		< .001
		3700 BLOCK TAYLORSVILLE RD		1		46		0.022		< .001
		BARDSTOWN RD AT I264		2		46		0.043		< .001
Note.  Proportions tested against value: 0.5.

The JASP software does allow sectioning out variables by value and so it was possible to narrow what happened there.  In this case the events at 3300 Bardstown were combined with 2200 Goldsmith Lane and I-264 since it is a unified business area.

Police Reports at 2200 Goldsmith Ln/3300 Bardstown Rd/I-264

Binomial Test
Variable		Level		Counts		Total		Proportion		p
CRIME_TYPE		ASSAULT		1		13		0.077		0.003
		DRUGS/ALCOHOL VIOLATIONS		2		13		0.154		0.022
		FRAUD		2		13		0.154		0.022
		MOTOR VEHICLE THEFT		3		13		0.231		0.092
		OTHER		2		13		0.154		0.022
		THEFT/LARCENY		3		13		0.231		0.092
PREMISE_TYPE		CONVENIENCE STORE		2		13		0.154		0.022
		CYBERSPACE		1		13		0.077		0.003
		DEPARTMENT / DISCOUNT STORE		1		13		0.077		0.003
		HIGHWAY / ROAD / ALLEY		2		13		0.154		0.022
		HOTEL / MOTEL / ETC.		2		13		0.154		0.022
		PARKING LOT / GARAGE		1		13		0.077		0.003
		RESIDENCE / HOME		1		13		0.077		0.003
		SERVICE / GAS STATION		1		13		0.077		0.003
		SPECIALTY STORE (TV, FUR, ETC)		2		13		0.154		0.022
Note.  Proportions tested against value: 0.5.

The reports were spread out among premise types and crime types.  There is no one pattern that appears to stick out.

I have anecdotal information that a car window was shot out on Lisbon but there was no police report in the database. 

The Bon Air Neighborhood Association continues to promote the idea of Neighborhood Watches.  They appear to be quite effective where they are active.  We will bring up the topic at our General Membership Meeting on Monday, March 9, 2020 at 6:30 pm at Buechel Park Baptist Church, 2403 Hikes Lane.  . 

Bon Air Housing Values: Rentals in Bon Air and Their Locations

This is the fifth in a series of posts about housing values in the historic Bon Air Neighborhood.as defined by the following boundaries: Furman Blvd on the east, Hikes Lane on the south, Bardstown Road on the East, and the Watterson Expressway/Taylorsville Rd on the North.

The database for this research was created from the Jefferson County Property Valuation Administrator (PVA) website and the Lojic.org website.  Both have similar data, but the PVA site could be pasted into an Excel file. The database was compiled between February 2019 and December 2019. The information was subject to some change based on property sales and re-evaluations.

The statistical software that is being used for the analysis is JASP which is an open source program created by the University of Amsterdam.

This particular post is subject to more interpretation given the data sources.  Understanding what single family homes precisely are rentals is not possible.  There were some houses that were obviously rentals as the owner was a limited liability corporation (LLC).  Some homes were owned by trusts that for the purposes of this study are being called "Unknown."   Early in the data collection in 2019, the PVA had open access to the detailed information that included the "contact address" for the property. (That access ended when reassessment time was over.)  

This particular variable was akin to what researchers call a "convenience sample" because some anecdotal evidence was used regarding known rental properties.  It was evident that the contact address was not the Bon Air property itself, which suggests that the property was a rental.  It is likely that there are more rental properties than could clearly be identified. 

Anyway, according to the database there are 229 properties or 11.6 percent of the single-family homes in the Bon Air Neighborhood boundaries that are either owned by LLC's or trusts.  For the record, one of the single-family homes on Hikes Lane is a business.  One was foreclosed at the time of the collection and its current status was not clarified. .    

The Bon Air zone is clearly the area with the most rentals with 139 properties that were either owned by LLC's or Trusts.  The Highgate Springs zone has 76 properties that are either rentals or owned by trusts.  The Bon Air Estates has 14 properties . 

The last subject to explore in this post is: What are the streets where the rentals are?  (If the table is a little small for your eyes, you can press [ctrl] and move your mouse wheel forward to zoom). Talisman Road in Highgate Springs and Terrier Lane in Bon Air tied for having the most rentals or possible rentals with 16. Meadow Lane (Bon Air), Stanton Blvd (Highgate Springs) , and Rio Rita Avenue (Bon Air) tied for second with 12. Doreen Way and Boaires Lane tied for third with nine.

Next week I will further discuss if there are any landlords who own multiple homes in the neighborhood. 

  

A Graphical Retrospective of 2019 Police Reports In the Historic Bon Air Neighborhood

The following are some graphs from police reports in the Bon Air Neighborhood for the Bon Air Neighborhood based on data posted by the Louisville Metro Police Department at https://data.louisvilleky.gov/dataset/crime-reports/resource/6d81fdf8-27e7-41ac-9ae4-1bfd8730f002   

As usual it includes the police reports within the boundaries of the historical Bon Air Neighborhood as defined:
1) Bardstown Road on the West  
2) Hikes Lane on the South
3) Furman Blvd on the East and

4) Taylorsville Rd and the Watterson Expressway on the North. 

It is a Microsoft Excel file that required manipulation and sorting.  Three relevant variables were created for month, zone and day of week.  All of the missing person reports for 2300 Goldsmith Lane were removed.  The edited Excel CSV file was then loaded into JASP, which is an open source statistical package.  This analysis does include events that happen on the edge of the neighborhood.

Not all of the results were practical to show.  There were 625 police reports.  There were too many different types of crime reports to show and too many different streets to show graphically. If requested,  I will be happy to create something with the software . . . if it is possible.

It was obvious that the 40218 portion had more of the crime reports. The next obvious result is that the Bon Air zone had the most reports.

August was the month with the most police reports in 2019. 

Monday was the day of the week with the most reports. 

Here is a list of where the police reports were at by premise or type of property.  You can draw inferences where the different reports were made. 

It is hoped that 2020 is a safer year in the Bon Air Neighborhood with fewer police reports.  The Bon Air Neighborhood Neighborhood Association is very much in support of neighborhood watches, especially in light of expected Metro Government budget cuts that will likely affect LMPD.  The more we engage in crime prevention, the more we allow Louisville's Finest to attend to higher priority situations. 

I hope that you can come to BANA's next general membership meeting on Monday, March 9, 2020 at 6:30 pm at Buechel Park Baptist Church. 

.


.

Bon Air Housing Values: Values by Zone

This is the fourth in a series of posts about housing values in the historic Bon Air Neighborhood.as defined by the following boundaries: Furman Blvd on the east, Hikes Lane on the south, Bardstown Road on the East, and the Watterson Expressway/Taylorsville Rd on the North.

The database for this research was created from the Jefferson County Property Valuation Administrator (PVA) website and the Lojic.org website.  Both have similar data, but the PVA site could be pasted into an Excel file. The database was compiled between February 2019 and December 2019. The information was subject to some change based on property sales and re-evaluations.

The statistical software that is being used for the analysis is JASP which is an open source program created by the University of Amsterdam.

The data set was adjusted one more time from last week to eliminate an apartment house did not get removed last week.

To cut to the chase,(and as illustrated in the box plot diagram)  housing values tend to be higher in the Bon Air Estates zone with a higher median (middle value) price of $141,800. It has a standard deviation of 28571, which means the majority of the values run between  and  113117 and 170259.

The Bon Air zone had the next highest median price with $135,000. It had a wider standard deviation of 30215 which means the majority of values run between $109656 and $170086.

The Highgate Springs zone has the lowest median price with $132,540.  The standard deviation was 26069 which means the values mainly ran between $108092 and $160230.

Nerd alert here for a moment.  The mode is the most repeating value in a set of data.  The mode in both Bon Air and Bon Air Estates is lower than the median (middle point of rank order) and mean (mathematical average). 

However, the most recurring value for Highgate Springs ($155,000) is way higher than the both the mean ($134161) and the median ($132,540).    Why is this? 

Maybe we should do one more thing here and create another chart called a "histogram" to look at how the values are spread out.

Here they are . . .

So, with these charts, we have to talk again about "skewness," which means distortion, but what pulls the numbers in a direction.

The Highgate Springs numbers are almost normally distributed but have a positive skewness (skewness value of 0.405), which could answer for why the mean and median are lower than than the mode. In real terms there are more higher house values that skew the center of values.  Nerd moment over.

To finish the skewness thoughts, the Bon Air numbers are positively skewed by a number of very high value houses.  The Bon Air Estates housing values are actually negatively skewed by a few lower values. 

Overall, the values reflect the diversity in housing stock.  Even though it was excluded from this data set, the oldest house in existence is the Bray Mansion (now the Bashford Manor Bed and Breakfast), followed by one of he original Hikes homes, and a few other older farm houses built prior to the development of the 1950's.   From 1952 to approximately 1957, a number of different developers built different models of houses at different price points for a number of different customer bases. 

While not perfectly homogeneous, the Bon Air Estates zone is the most consistent in having houses with larger square footage and basements. This reflects the higher median value.

The Highgate Springs zone has more diversity in housing stock than Bon Air Estates, but less than Bon Air.  There are blocks with what could be called tract or identical houses.  Some of these blocks are of higher value that include being brick houses with basements.  There are a number of blocks that are the the California Ranch (or kit houses) and the US Steel prefabricated houses (basic three-bedroom ranch houses on concrete slabs).

The Bon Air zone has the most diverse tracts with a variety of house models and values.  The higher valued houses skew the overall numbers. 

As I have written this post, I have had good memories from my time in J.Allen Whitt's class at the  University of Louisville learning SPSS and regression analysis.  Dr Whitt was also one of my dissertation advisers.  I have gratitude for his tutelage in understanding statistics and writing up research.   










.

Bon Air Housing Values: A Second Look at Overall Values With Better Software

This is the third in a series of a look at housing values in the Bon Air Neighborhood as defined by the following boundaries: Furman Blvd on the east, Hikes Lane on the south, Bardstown Road on the East, and the Watterson Expressway/Taylorsville Rd on the North.

I confess that as I have written this post, it is the kind of stuff that is more egghead and boring and I hope that I do not sound too much like that weird guy you hated because he was actually excited about statistics class and he wrecked the bell curve.

The database for this research was created from the Jefferson County Property Valuation Administrator (PVA) website and the Lojic.org website.  Both have similar data, but the PVA site could be pasted into an Excel file. The database was compiled between February 2019 and December 2019. The information was subject to some change based on property sales and re-evaluations.

The statistical software that is being used for the analysis is JASP which is an open source program created by the University of Amsterdam.

The dataset was adjusted to eliminate non-housing properties that reduced the data set from 1995 cases (used in the last post) down to 1975 cases (or properties with houses on them).

An initial run of descriptive statistics was run with the following results:

Statistical Table with Min and Max

Descriptive Statistics
		VALUE
Valid		1975
Missing		13
Mean		138316.319
Median		135000.000
Mode		155000.000
Std. Deviation		30185.350
Skewness		1.555
Std. Error of Skewness		0.055
Kurtosis		11.110
Std. Error of Kurtosis		0.110
Range		468400.000
Minimum		2110.000
Maximum		470510.000
25th percentile		118650.000
50th percentile		135000.000
75th percentile		153900.000

The most valuable numbers out of this table are the Mean (average) and Median (middle point of the numbers).  For Bon Air as a whole the mean or average home value was $138,216.  The Median was 135,000.  This initially would say that matters are an even deal, however, when you look at graphics--especially the bar graph at the beginning of this post, it is very skewed or distorted. to the higher values with a value of 1.555 (should be 1 or less).

The skewness makes some sense when you see a range of $470510 (distance between highest and lowest number.  The maximum home value was $470510 and the lowest privately held property is $2110.  The highest valued property is a bed and breakfast and the lowest value is a piece of that triangle/cemetery at Bon Air and Goldsmith Lane (that one crooked part).

 So, an adjustment was made by taking out the highest and lowest properties and re-running the analysis. Here are the results

Adjusted Descriptive Statistics

Descriptive Statistics
		VALUE
Valid		1973
Missing		13
Mean		138216.984
Median		135000.000
Mode		155000.000
Std. Deviation		29098.145
Skewness		1.045
Std. Error of Skewness		0.055
Kurtosis		4.487
Std. Error of Kurtosis		0.110
Range		315620.000
Minimum		5000.000
Maximum		320620.000
Sum		2.727e +8

Cutting out the highest and lowest value did fix the skewness or distortion and it was significant that the mean (average) median (middle) values stayed virtually the same. Of course the range was shortened down to $315,620 with a high value of $320,620 and a vacant property value of $5000.

At this point that silly term "standard deviation" might be of some value.  It is a group of numbers that cluster around the average or mean.   Before adjustment the standard deviation was  $30,185.  After adjustment the standard deviation was $29,098.  It changed less than $1000.  Before I lose anyone else, let's look at a graph called a "Box Plot."

The box plot is valuable in illustrating the standard deviation in this case. It shows the majority of the housing values cluster between approximately $130,000 and $170,000.  That is the middle of the pack.  If we were to be precise, the central tendency of the housing values in the neighborhood are between $149,549 and $120451. The cluster appears to be just bigger than the standard deviation. There are a few outliers that still skew the data but not as much.

The next planned post in this series will be about the values of houses in the different zones.  It might explain some of the overall housing values.

We in the Bon Air Neighborhood Association are about doing things that improve our housing values and that includes our current arts development initiative.  We hope that you will join us as a member if you are not already one.  Our next meeting is tomorrow, Monday February 10, 2020 at 6:30 pm in the Bon Air Library on the "Boulevard."