Forecasting
Management Report
Forecasting
Management Report
Table of Contents
Table of Contents 0
Introduction 2
Problem 2
Objective 3
Methodology 3
Analysis and Results 4
Question 1: 4
Question 2: 5
Price vs Neighbourhood 5
Price vs Special Feature 6
Price vs Pool 7
Question 3: 8
Question 4: 9
Question 5: 13
Question 6: 13
Question 7: 14
Question 9: 15
Question 10: 16
Question 11: 17
Question 13: 18
Price vs Building size and Year Sold 18
Price vs Number of Bathroom(s) and Bedroom(s) 18
Conclusion: 19
Introduction
The Vancouver housing market has seen a dramatic rise in prices over the past years. This varies with the property type as well. The overall objective of this research is to identify the price of houses sold in North Vancouver and the factors that affect it. To conduct this research, stats will be involved with the data provided.
When assessing markets in a foreign district many different variables go into the decision making. To help evaluate relationships between needs and wants of a new property we use statistical data to display qualitative and quantitative results to assist buyers in making the right decision. An Evaluation of key factors to predict and show housing prices in North Vancouver based on the sample size of 100 houses that were sold within the last 10 years gives us an accurate reflection of trends and features that create more cost for buyers.
Problem
The housing market in Vancouver is more expensive than the market in Ottawa. Below are the key indicators we will be comparing in the report
Year Sold
Property Type
Age
Building Size
Lot Size
Neighbourhood
Number of Bedrooms
Number of Bathrooms
Special Features
Pool (Yes/No)
Price Sold
If the market proves to be too expensive David will not be able to move to North Vancouver.
Objective
We are conducting a research in the housing market for North Vancouver, due to the city being more expensive than Ottawa and many other cities. As a realtor in BC we have created a brief report on the price of the houses sold in North Vancouver over the past 10 years. We decided to use a sample size of 100 in order to prepare this report for current and future potential buyers in North Vancouver.
Methodology
Our data suggests that the average price per home in the 100 samples is $1,212,330.
Figure 1 – Appendix A
*Note: There were 3 outliers in the above data 16, 57, 94. These were included in all of the displayed graphs.
Throughout the report, the reader will see various data sets to help make informed and educated conclusions. These sets of data include; age, year sold, location, number of bathrooms. We decided against using several bedrooms, lot size, and building size as indicators of the current market situation.
Qualitative and quantitative data have been displayed to help explain the current market. To conclude the data, multiple regression analyses, observational errors, sampling variations, mean and standard deviations were used. These techniques allowed us to predict trends in the data and allowed us to observe relationships between data sets.
Analysis and Results
Question 1:
Has the mean price varied over the years?
The mean price has varied over the years with a large peak in 2015. 2019 was the most expensive year for average prices in North Vancouver.
Figure 2 – Appendix B
Question 2:
Use appropriate charts to describe individually the relationship between the key response variable (price sold) and the following key explanatory variables (neighborhood, special feature, pool). Comment on the result.
Price vs Neighbourhood
Figure 3 – Appendix C
Each Neighborhood had a minimum of 4+ samples to give us an accurate average for prices in each neighborhood excluding Lions Gate – this neighborhood only had 2 samples that were very different – creating a very unrealistic average. Please exclude the Lions Gate Neighborhood from the analysis.
The data indicates that Deep Cover is the most expensive area to purchase. The cheapest option would be Seymour.
Price vs Special Feature
Figure 4 – Appendix D
The data suggest properties with a Waterfront are the most expensive. The difference between Water view and No special feature does not have a significant delta.
Price vs Pool
Figure 5 – Appendix E
Having a pool as a feature in a property will affect the price. This will depend on the value a customer sees in having a pool at home.
Question 3:
Construct a pivot table to group the following variables: Neighborhood vs. special feature. Use various statistics tools to describe the relationship between variables. Comment on the result.
Figure 6 – Appendix A
**The pivot table above will allow the reader to filer between special features and Neighborhood, unfortunately since this is a copy-paste you will not be able to filter.
This pivot table shows the different neighborhoods, and comparisons between having a waterfront, water view or ones that do not have either.
According to the data, the homes in Deep Cove with a waterfront are the most expensive among the samples. Upper Lonsdale has the highest density of properties with a water view.
Lynn Valley has neither properties with a water view or a waterfront.
Question 4:
Which of the variables (year sold, age, building size) is the most and least highly correlated with price?
Below are the charts for the variables of the year sold, age, and building size. Please see question 5 for analysis.
Variables for Building Size
R-Square- 0.033815081
Variables for Year Sold
R- Square – 0.0315523367995543
Variables for Age
R-Square – 0.000265818784686657
Multi Regression Scatterplot Graph
Figure 7 – Appendix F
Figure 8 – Appendix F
Figure 9 – Appendix F
Figure 10 – Appendix F
Figure 11- Appendix F
Figure 12 – Appendix F
The charts below show the regression of each variable with a y-axis equal price.
Variables for Building Size
R-Square- 0.033815081
Figure 13- Appendix K
Variables for Year Sold
R- Square – 0.0315523367995543
Figure 14 – Appendix L
Variables for Age
R-Square – 0.000265818784686657
Figure 15 – Appendix J
Variables for Building Size
R-Square- 0.033815081
Figure 16 appendix – K
Question 5:
According to the regression analysis that was referenced through the data provided, the results indicate that the highest correlation is building size with an R-value of 0.033815081 and the least being the age with the R-value of 0.000265818784686657. Both are not significantly correlated.
This Photo by Unknown Author is licensed under CC BY-SA-NC4
This Photo by Unknown Author is licensed under CC BY-SA-NC4Question 6:
Looking at the regression analysis that was extracted from the data it displays that price and the age of the house are not strongly correlated with an R-value of 0.033815081.
Question 7:
Is there a relationship between the # of bathrooms and price sold?
Figure 17 – Appendix A
The relationship between the number of bathrooms and the price a house is sold is not strongly correlated. Other factors pose a stronger relationship.
R-Square: 0.00291567208922046
Figure 18 – Appendix L
Question 9:
Does having a pool matter? Is there a difference in the mean and distribution of price between the two groups (pool versus no pool)?
Figure 19- Appendix A
The above chart indicates that there is a significant difference in price with and without a pool such as having a pool. The cost of having a special feature versus no special feature is $568,923 or a 40.97% difference.
Median
No Pool: $830,000
With a Pool: $1,300,000
Based on the data set, we have attended binomials by introducing binary of zero and one, but we still got the error “Regression-input range contains non-numeric data”. Therefore we will eliminate these variables since it’s not numerical.
Question 10:
Does having a special feature matter? Is there a difference in the mean and distribution of price between the two groups (special feature versus no special feature)?
Figure 20 – Appendix G
The above chart indicates that there is a significant difference in price with and without a special feature such as having a waterfront or water view. The cost of having a special feature versus no special feature is $587,256 or a 41.59% difference.
Question 11:
Is there a difference in price for the different neighborhoods? Compare the means for the different neighborhoods.
Figure 21 Appendix C
As mentioned previously, based on the data collected on the 100 properties sold in the last 10 years in North Vancouver, Deep Cove had the highest mean of 2,099,375, while Lion’s gate had the lowest at 637,500.
When looking at feasible choices, the neighborhood can be broken down into four categories. Each category represents similar pricing with a distribution of mean no greater than 20% between the highest and lowest price among the different neighborhoods.
Option 1: Lions Gate, Seymour, and Upper Capilano with average properties between
$637,500 – $760,000
Option 2: Lower Lonsdale, grand Boulevard with average properties between
$916,083 -$975,833
Option 3: Central Lonsdale, Westview. Lynn Valley, Upper Lonsdale with average property prices between $1,126,563 – $1,355,682
Option 4: Deep Cove with a mean of $2,099,375
Question 13:
Create the best multiple regression model you can predict price.
Price vs Building size and Year Sold
Average(s)
ERROR
MAD
MAPE
$ 262,574.76
31%
$112,977,247,222.76
Figure 22 Appendix H
When comparing price versus year sold and building size the multiple regression gave an error of 41%. By eliminating the outliers with the ID numbers 16, 57, and 94, the error decreased by 10 percent down to 31%. Unfortunately, this error is still extremely high, making it difficult to forecast successfully based on these two variables.
Price vs Number of Bathroom(s) and Bedroom(s)
ERROR
MAD
MAPE
$ 292,573.60
35%
$ 129,799,111,323.33
Figure 23 Appendix I
When comparing price versus year sold and building size the multiple regression gave an error of 35%. This result does not include the outliers with the ID numbers 16, 57, and 94. Unfortunately, this error is still extremely high, making it difficult to forecast successfully based on these two variables.
Conclusion:
The mean price has varied from 2015 – 2019. There was a large peak in 2015 due to an outlier from home ID #57. Market trends over the year make 2019 the most expensive year. Based on these findings, purchasing sooner than later is in the best interest of the buyer. When looking at neighborhoods we found a deep cove to be the most expensive. The data supplied for the lion’s gate neighborhood was very limited (2 samples) thus making it difficult to forecast on property value.
Our findings showed properties with waterfront were significantly more expensive than the rest. Having a pool tends to increase property prices.
When helping a homeowner determine value, building size had the strongest correlation with price. Whereas, the age of the house had the weakest correlation with price.
When comparing bathrooms to price, we saw homes with 4 bathrooms to be more expensive than houses with 5 or more bathrooms. Unfortunately, this was due to properties with 4 bathrooms being in locations with higher property value such as Upper Lonsdale.
The relationship with houses that have pools and non-pools is drastic, we want to inform buyers that there is an average increase of 40% in the price of the home if they choose to buy a property with a pool.
Overall, customer preference plays a huge role in the purchase of a new home. Location, size, age, and features will affect the price – a customer’s budget will be a large indicator as to whether or not they will be able to afford a specific property.
19 | Page