Statistical Analysis Plan (SAP) 范例

---

@author: Mingran Jia

• URL of data:
  https://nethouseprices.com/house-prices/Lanarkshire/GLASGOW?page=1
  https://nethouseprices.com/house-prices/Lanarkshire/GLASGOW?page=2
  …
  https://nethouseprices.com/house-prices/Lanarkshire/GLASGOW?page=10

• The context of the data:

  > glimpse(house_price_glasgow)
  Rows: 500
  Columns: 3
  $ address <chr> "1 Ettrick Place, Glasgow, G43 1UA", "2~
  $ prices  <dbl> 144500, 212750, 185000, 90000, 126894, ~
  $ types   <chr> "Flat", "Flat", "Flat", "Semi Detached"~
  

• The content of the data:

  > str(house_price_glasgow)
  tibble [500 x 3] (S3: tbl_df/tbl/data.frame)
   $ address: chr [1:500] "1 Ettrick Place, Glasgow, G43<U+00A0>1UA" "2/1 26 Tassie Street, Glasgow, G41<U+00A0>3QF" "Flat 1/3 18 Prospecthill Grove, Glasgow, G42<U+00A0>9LD" "41 Ochil Street, Glasgow, G32<U+00A0>7SD" ...
   $ prices : num [1:500] 144500 212750 185000 90000 126894 ...
   $ types  : chr [1:500] "Flat" "Flat" "Flat" "Semi Detached" ...
  

---

Research Questions

We wish to evaluate the relationship among house types, locations and prices to assist real estate developers set more reasonable price.

• The impact of different housing types on house prices
• The impact of different regions on house prices
• The mutual influence of different areas and room types

---

Statistical Analysis Plan for House Price

Population

• Glasgow House Price Statistics

Primary Objective:

• Estimate the influence of house types and locations on house prices

Secondary Objectives:

• Assess the top-heated house types and postcodes in the city
• Estimate the mutual influence of different locations and room types

Data Collection methods:

• Scrap the most recent 500 house prices of 201191 total in Glasgow from the house saling website of Scotland as the sample to represent the population of Glasgow house statistics.
• The house is identified by an unique address.
• The house is classified by limitedly different types.

Variables Under Consideration:

• House prices grouped by area; house prices grouped by different type; difference in house price for different type in the same area; difference in house price for different area in the same type - Primary outcome variable
• Areas division accessed by locations - Primary explanatory variable
• Top house type; top house area; top and bottom selling house type in each area; top and bottom selling house area of different types - Explanatory outcome variable

Missing Data Procedures:

• If any data of the house type or location is missing, that house is excluded from analysis.
• If the price is missing, use the average price of that area; if there are less than two house prices in that area, then that area is excluded from analysis.

Summaries to be presented:

• Basic statistical discription applied to house price including mean, standard deviation, median, etc…

Models to be fitted

• Linear model to be