Introduction

This study is intended to be submitted as a Final Project for Professor Cline’s Data Visualization class. The following study is an attempt to analyze the impact of COVID-19 on housing prices in the UK with an emphasis on diagrams, plots and interactive visualizations. As an international student from England, I have many questions regarding the impact of COVID-19 on English communities. Whilst I am interested in the impact of COVID-19 on housing prices, I am also interested if housing prices has an impact on COVID-19. For example, communities with cheaper housing and less funding may receive less healthcare privileges and have a high proportion of people dying from COVID-19. Despite my interest, my research will not cover the converse and will be instead focusing on the effect on COVID-19 on housing prices.

The data started off messy. The data I found was the only reliable sources of data I could find for both housing and COVID-19 data. One was partitioned into counties whilst the other was partitioned into unitary authorities and counties. This mismatch between the data sets lead me to do lots of filtering and cleaning to facilitate the joining of the data sets, hence why some counties appear multiple times on some maps and others are missing entirely. For my study, I decided to use median house prices to reduce the effect of some counties with significant outliers and few observations, such as the City of London, and the mean number of new COVID-19 cases a day.

I expect to see in my project that an increase in COVID-19 cases causes an area to be less desirable to live and leads to a decrease in housing prices.

Sources

COVID-19 data was taken from https://coronavirus.data.gov.uk/details/download, the official UK Government website for data on COVID-19. The data contains date and county information alongside death, cases and vaccination statistics. Housing related data comes from https://www.gov.uk/government/statistical-data-sets/price-paid-data-downloads, again the official UK Government website for data on housing. The dataset provide information on the location, date and price of house sales in the UK. I retrieved county boundary data from The Office for National Statistics (https://geoportal.statistics.gov.uk/search?collection=Dataset&q=County&sort=name).

Location of Counties

Brief Overview of The Relationship of Median House Price and Different COVID-19 Variables

Across all COVID-19 variables there seems to be a positive relationship between median housing prices. An interesting find is that the more people vaccinated in a county, the higher the median house price of the county. This relationship could be due to these counties being more educated and “privileged” which is associated with higher vaccination rates alongside the shown increase in housing prices. It could also be due to counties having higher populations and therefore logically more people vaccinated and more urban and populated areas are desirable to home seekers and increase house prices. Despite this, the causality of the remaining plots remain ambiguous. For research and time purposes, I will be focusing on the effect of COVID-19 cases.

Scatterplot of Mean Number of New Cases a Day and Median House Price of Each County By Month

This plot is the same as above but its enlargement allows for easier interpretation. There is a strong positive relationship between Median House Price and Mean Number of New COVID-19 Cases a Day. Whilst this relationship does not appear strong shown by the linear model, the linear model is heavily affected by outliers (>2000 cases a day) which force the linear model to towards a more negative relationship.

Each dot represents the mean of both housing prices and new COVID-19 cases a day in a county for a month. I have shown the aggregate below used to form this plot to assist further understanding.

head(joined2)
##                                county year month  price newCasesByPublishDate
## 1        BATH AND NORTH EAST SOMERSET 2020    01 330000                    NA
## 2                             BEDFORD 2020    01 270950                    NA
## 3               BLACKBURN WITH DARWEN 2020    01 110000                    NA
## 4                           BLACKPOOL 2020    01 110000                    NA
## 5                       BLAENAU GWENT 2020    01  78000                    NA
## 6 BOURNEMOUTH, CHRISTCHURCH AND POOLE 2020    01 275000                    NA

Linear Model

##                    term     estimate  std.error  statistic      p.value
## 1           (Intercept) 228905.74363 2220.36287 103.093844 0.000000e+00
## 2 newCasesByPublishDate     49.45003   10.36045   4.772963 1.933637e-06
##       conf.low    conf.high   df outcome
## 1 224551.53630 233259.95095 2218   price
## 2     29.13284     69.76722 2218   price

The positive relationship is also presented running a linear model which shows for each increase in mean number of COVID-19 cases a day, the median house price increases by $49.45. This increase is most likely not because of the change in number of COVID-19 cases, instead it is probably caused by more populated areas being desirable places to live which drives up housing prices whilst also being more susceptible to COVID-19 outbreaks.

Tap on dots to reveal county and other points with the same county

It appears the variation in median house price is due to changes in counties rather than COVID-19 cases because dots of a county appear around the same area on the y axis as other dots with the same county (highlighted in yellow when hovering over a dot). The increase in COVID-19 cases in a county does not appear to be having an affect on the median house prices in that county.

Comparison of Median House Price and Mean Number of New Cases a Day by County using Chloropleths

Tap on counties of each chloropleth to compare values

Chloropleth using Color and Size of Bubbles to Represent Median House Price and Mean Number of New COVID-19 Cases A Day for Each County

Tap on each county to observe values

Relationship of Mean Number of New COVID-19 Cases a Day and Median House Prices for Each Month

I recognize that conducting analyze across counties without a population density variable is causing my results to become inaccurate. The above info graphic supports the claim that the increase in median housing prices is not caused by the increase in COVID-19 cases. The spread of median housing prices across the y axis remains constant throughout each month. The strong positive linear relationship arises in winter months such as December and January when COVID-19 cases are at its highest. Hence, I have decided to narrow my research into my home county East Sussex to remove the affect of population density disparities between counties

Conducting Further Analysis into The County of East Sussex

Click on a point to distinguish its matching point in opposite plot

The same relationship is present when only narrowing down to research area to East Sussex. This suggests that whilst population densities disparities across counties is playing a large factor in the increase in housing prices, it is not only factor affecting housing prices.

Conclusion

A brief search online (https://www.forbes.com/uk/advisor/personal-finance/2022/04/14/house-prices-updates/) makes it clear that housing prices during the COVID-19 pandemic have increased by an estimated 10.9% across the UK. Articles state that this increase in housing prices in caused by an increase in demand alongside a diminished supply. Such as factors, including population density, economic climate and demand for houses in urban areas makes it hard to see any accurated relationship between COVID-19 and housing prices. In conclusion, no reliable and sensible statement can be made about the relationship between COVID-19 cases and housing prices from my study. The only logical discovery found was that counties with more people vaccinated tended to have higher house prices.