IT 110 project
AGKT Solutions Presents: The Bauer Project
Manas Aryal
Michael Gomes
Mohamed Kholandi
Nathaniel Trippe
About Us
Mr. Bauer, thank you for considering AGKT to assist you with your next real estate investment.
AGKT Solutions, Inc. is an analytic consulting firm based in Pittsburgh, PA and the surrounding metropolitan area
We specialize in data-driven decision making and investment advising
Our goal is to give you and your family the most informed and logical option for choosing your next home: No personal preference, no bias, just the truth.
Our Methods and Practices
Gathering: We collect statistics from trusted sources to form our decisions, removing emotion from our recommendations
Processing: Data mining and modeling using advanced software such as RapidMiner to sort out over 3,000 prospective properties in the city of Pittsburgh
Predicting: The most important criteria determines our model for real estate recommendations – crime rate, price, neighborhood proximity to goods and services, traffic and accessibility
You have protected us for so many years – it is an honor for us to help return the favor and protect you and your family.
The Process
Manually labeled 201 properties – used as training data
Training data includes at least one entry from all 90 Pittsburgh Neighborhoods
Our rating system grades properties from 1 (lowest quality) to 10 (highest quality) – based on real data on pricing, quality of neighborhood, and crime rate
Traffic data used in post-process decision making to help differentiate highest-rated properties
Criteria
Price based on budget request of $500,000; the cheaper, the better, but we wanted to include room for slightly above budget if there is excellent value on the property
Crime ratings, taken directly from City of Pittsburgh Police Bureau website, are based on an average of crime rates from all neighborhoods – a few were removed as their rates were too high
This can be explained by neighborhoods that are very small in population
Average crime rate/neighborhood came out to roughly 13 crimes per 100 people, with outliers removed.
Neighboorhood based on a list compiled by online source, RoadSnacks.net, that uses “High unemployment, Low median income, Low population density, Low home values, High crime” as criteria for low-quality neighborhood.
Worst neighborhood is ranked “1” in list; best is “90”.
Criteria
| Price Criteria | RATING | Crime Data Criteria | RATING | Neighborhood Criteria | RATING |
| $1-$300,000 | 8-10 | 0-6.5 Crimes per 100 people | 8-10 | 71-90 | 9-10 |
| $300,001-$500,000 | 6-7 | 6.51-13 Crimes per 100 people | 5-7 | 51-70 | 7-8 |
| $500,001-$600,000 | 4-5 | 13.01-19 Crimes per 100 people | 3-4 | 41-50 | 5-6 |
| $600,001-higher | 1-3 | 19.01-higher | 1-2 | 1-40 | 1-4 |
Below: Cover of 2014 City of Pittsburgh crime report and sample data
Above: Sample of real estate data with price listings
Below: Sample of neighborhood rankings from RoadSnacks.net
The Process…
Modeling and Predicting
Decision Tree: Our manually labeled “training data” was used to model ratings for remaining unrated property entries (the “test data”)
Decision tree determined rules for how the properties were rated.
We added a binomial attribute, with “buy?” as label, to determine “yes” (buy) or “no” (don’t buy).
Our decision tree determined that any property 7 or greater, with certain rule exceptions, is classified as a “yes”.
Modeling and Predicting
Below: Decision Tree
Longitude less than or equal to -79.889 > AVERAGE/OVERALL = 9 > yes
Longitude less than or equal to -79.889 > AVERAGE/OVERALL = 8 > Price less than or equal to 266,666 > yes
Longitude less than or equal to -79.889 > AVERAGE/OVERALL = 7 > Price less than or equal to 84,750 > no
Longitude less than or equal to -79.889 > AVERAGE/OVERALL = 5 or 6 > no
Modeling and Predicting
Data was then processed to give numerical predictions between 1-10
We had a total of 533 properties rated 9 or better
Only 7 were rated at 9.3 (rounded), our highest rating
Below: Prediction results in RapidMiner
Modeling and Predicting
We then ran another model to determine the accuracy of our predictions…
Modeling and Predicting
…And our result was 76.18% Accurate:
Our Final Determination
Our ratings included traffic data after the RapidMiner processes completed
Data taken directly from shinyapps.io, powered by Google Maps
Map represents relationships between traffic patterns and real estate values
Pictured: Visualization of traffic data by neighborhood. Ratings go from dark red (lowest value) to dark green (highest value).
Our Final Determination
However, all 7 of our highest rated properties are from the same neighborhood – Point Breeze, so this criteria did not narrow down our results any further.
Point Breeze is our highest rated neighborhood in the entire City of Pittsburgh
Our final criteria for the 7 properties, done outside of RapidMiner processing, was highest current valuation in USD.
And the winner is…
315 East End Avenue, Pittsburgh, PA
315 East End Avenue, Pittsburgh, PA
315 East End Avenue has seen increase in value from $242,000 in 2014 to an estimated $432,765 – Just over a $190,000 return on investment, or roughly a 78% increase.
3-bedroom, 3.5 bath, 2,496 square feet
Only 1.8 miles away from Taylor Allderdice High School, 0.7 miles away from Pittsburgh Sterrett 6-8 Middle School
1 minute drive and 0.2 miles from LifeCare Hospitals of Pittsburgh
3 minute drive and 0.6 miles to Frick Park – the largest municipal park in the City of Pittsburgh, covering over 644 acres.
315 East End Avenue, Pittsburgh, PA
Not to mention, Point Breeze has a crime rate of only 4.05 crimes per 100 people per year – one of the lowest in Pittsburgh.
We see this as a property that hits on the most crucial points: In a safe, convenient neighborhood with a lot of services to offer, and that is under your requested budget of $500,000.
We thank you for your time and wish you the best of luck in your search for a new home!
Sources and Citations
Crime Data: Peduto, William, Mayor, and Cameron McLay, Chief. "City of Pittsburgh Department of Public Safety, Bureau of Police, 2014 Statistical Report." (n.d.): n. pag. City of Pittsburgh. City of Pittsburgh, 2014. Web. 2 Nov. 2016. <http://apps.pittsburghpa.gov/pghbop/2014_Annual_Report_Final_draft.pdf>.
Neighborhood Rankings: Marksman, Frederick. "These Are The 10 Worst Pittsburgh Neighborhoods." RoadSnacks. N.p., 8 Mar. 2016. Web. 4 Dec. 2016. <https://www.roadsnacks.net/worst-pittsburgh-neighborhoods/>.
Traffic Data: Mapbox, and Leaflet, eds. "Relationships between Traffic Patterns and Real Estate Values from March to August 2014; Pittsburgh, PA." Shinyapps.io. N.p., n.d. Web. 11 Nov. 2016. <https://vietexob.shinyapps.io/traffic_real_estate/>.
Sources and Citations
Jack Bauer Image #1: Ryan, Christopher, Ph.D. Jack Bauer 24 Image. Digital image. Psychology Today. Courtesy of Twentieth Century Fox, 11 Dec. 2014. Web. 4 Dec. 2016. <https://www.psychologytoday.com/blog/sex-dawn/201412/did-24-prime-the-pump-torture>.
Jack Bauer Image #2: Litle, Larry. Jack Bauer 24 Image. Digital image. Geekycool.com. N.p., 1 Jan. 2016. Web. 4 Dec. 2016. <http://www.geekykool.com/fox-orders-new-24-series-without-jack/>.
Pictures of Home: Zillow. "315 E End Ave, Pittsburgh, PA 15221." Zillow. Author Unknown, n.d. Web. 1 Dec. 2016. <http://www.zillow.com/homedetails/315-E-End-Ave-Pittsburgh-PA-15221/11341515_zpid/>.
Sources and Citations
Data on Homes/Final Comparison: Zillow, Inc. "315 E End Ave, Pittsburgh, PA 15221." Zillow: Real Estate, Apartments, Mortgages & Home Values. Author Unknown, n.d. Web. 1 Dec. 2016.
Zillow, Inc. "7715 Tuscarora St, Pittsburgh, PA 15221." Zillow: Real Estate, Apartments, Mortgages & Home Values. Author Unknown, n.d. Web. 1 Dec. 2016.
Zillow, Inc. "7718 Waverly St, Pittsburgh, PA 15221." Zillow: Real Estate, Apartments, Mortgages & Home Values. Author Unknown, n.d. Web. 1 Dec 2016.
Zillow, Inc. "7731 Edgerton Ave, Pittsburgh, PA 15221." Zillow: Real Estate, Apartments, Mortgages & Home Values. Author Unknown, n.d. Web. 1 Dec 2016.
Zillow, Inc. "7725 Edgerton Ave, Pittsburgh, PA 15221." Zillow: Real Estate, Apartments, Mortgages & Home Values. Author Unknown, n.d. Web. 1 Dec 2016.
Zillow, Inc. "550 Peebles St, Pittsburgh, PA 15221." Zillow: Real Estate, Apartments, Mortgages & Home Values. Author Unknown, n.d. Web. 1 Dec 2016.
Zillow, Inc. "583 East End Ave, Pittsburgh, PA 15221." Zillow: Real Estate, Apartments, Mortgages & Home Values. Author Unknown, n.d. Web. 1 Dec 2016.
Sources and Citations
Information on Frick Park: Wikipedia. "Frick Park." Wikipedia. Wikimedia Foundation, n.d. Web. 6 Dec. 2016.
Map Data: Google, DigitalGlobe. 2016. 315 East End Ave, retrieved from: https://www.google.com/maps/place/315+East+End+Ave,+Pittsburgh,+PA+15221/@40.4443802,-79.8968446,17z/data=!3m1!4b1!4m5!3m4!1s0x8834eddd24e485cd:0x31a84535a5b39cc4!8m2!3d40.4443802!4d-79.8946559