places_lived.Rmd

---
title: "Places I have lived"
# author: "Stuart Leeds"
date: "2021-07-14"
output: 
  html_document:
    toc: FALSE
---

```{r setup, include=FALSE}

knitr::opts_chunk$set(
  echo = FALSE, # set to TRUE to allow code chunks to "print" in output
  message = FALSE,
  warning = FALSE,
  fig.align='center'
)
```

> It occurred to me that I have lived in many places during my lifetime. This is
> a basic run-down of them all, giving rough locations as coordinates. I might
> return to it at some point and add dates, but this will do for now.

[`Code` available
here](https://github.com/sleeds50/my_website/blob/main/places_lived.Rmd)^[Feel
free to use it and change details to suit your needs - or improve it (if you do,
please email me to say how you did it better - we're all learning, right?)]

```{r libraries}

# load libraries
library(tidyverse)
library(maps)
library(mapproj)
library(leaflet)
library(kableExtra)
library(ggrepel)
```

# 1. Locations

Table 1 lists the places I have lived. Let me state that it only includes places
where I have lived with parents as a child and young adult (see items 1:8); and
places that I have officially rented or mortgaged as an adult (items 9:end).
Furthermore, if memory serves, we lived in two properties in Kuwait, two in
Bahrain; and I lived in Chertsey more than once; and rented twice in Addlestone.

I have not included sofa-surfing or temporary dwellings.

Table 1 also shows the order of properties lived in, from birth to now
(some 50-ish years later!)

```{r dwellings}

# create data frame
dwellings <- data.frame(
  List = 1:18,Address = c(
    "Abridge, Essex",
    "Stansted, Essex",
    "Kuwait City, Kuwait",
    "Manama, Bahrain",
    "Pilgrims Hatch, Essex",
    "Chertsey, Surrey",
    "Stradishall, Suffolk",
    "Cowlinge, Suffolk",
    "Addlestone, Surrey",
    "Twickenham, Middlesex",
    "Twickenham, Middlesex",
    "Bedfont Green, Middlesex",
    "Jaywick, Essex",
    "Harwich, Essex",
    "Harwich, Essex",
    "Ipswich, Suffolk",
    "Bath, Somerset",
    "Bath, Somerset"
  ),
  Latitude = c(
    51.651447,
    51.9070471,
    29.3797091,
    26.2235041,
    51.6357161,
    51.378971,
    52.139740,
    52.158376,
    51.371280,
    51.438541,
    51.434900,
    51.450761,
    51.776301,
    51.933189,
    51.932645,
    52.068666,
    51.394750,
    51.392352
  ),
  Longitude = c(
    0.124044,
    0.2017505,
    47.9735629,
    50.5822436,
    0.2945186,
    -0.520269,
    0.516622,
    0.506907,
    -0.489646,
    -0.336711,
    -0.332810,
    -0.439600,
    1.119079,
    1.274589,
    1.275031,
    1.130408,
    -2.388744,
    -2.384639
  )
)
```

```{r main table, }

# print a table to see the information at a glance, instead of having the reader try to 
# find and link each row in the format above.

dwellings |>
  kbl(caption = "<b>Table 1</b><br> <i>Places I have lived</i>") |>
  kable_classic_2(full_width = FALSE, position = "center")
```

```{r mean coords}

# These next two code-chunks create the variables for the mean and median coordinates, 
# which come in handy later on when we plot them.

# mean coordinates used for world map and table 2
meanCoords <- data.frame(
  row.names = c(
    "Worldwide",
    "England"
  ),
  Latitude = c(
    mean(dwellings$Latitude),           # World
    mean(dwellings[-c(3:4), ]$Latitude) # England
  ),
  Longitude = c(
    mean(dwellings$Longitude),           # World
    mean(dwellings[-c(3:4), ]$Longitude) # England
  ),
  Place = c(
    "Génicourt-sur-Meuse, France",
    "Cheshunt, Hertfordshire, England"
  )
)
```

```{r median coords}

# median coordinates used for world map and table 3
medianCoords <- data.frame(
  row.names = c(
    "Worldwide",
    "England"
  ),
  Latitude = c(
    median(dwellings$Latitude),           # World
    median(dwellings[-c(3:4), ]$Latitude) # England
  ),
  Longitude = c(
    median(dwellings$Longitude),           # World
    median(dwellings[-c(3:4), ]$Longitude) # England
  ),
  Place = c(
    "Upminster, Essex, England",
    "Stapleford Abbotts, Essex, England"
  )
)
```

# 2. Interactive map

The coordinates are pin-pointed on a world map (use +/- zoom functions and 
left-click[hold]/drag to pan). Hover cursor over circles for labels.

* __Green circles:__ points of all places lived.

* __Blue circles:__ _mean_ average location of ALL coordinates in _Table 1_ 
(see _Table 2_ for details)

* __Red circles:__ _median_ average locations of ALL coordinates in _Table 1_ 
(see _Table 3_ for details)

```{r map}

# plot coordinates on the world map
# whole table
worldMap <- with(dwellings, leaflet(width = 900) |>
  addTiles() |>
  addCircleMarkers(
    lng = Longitude,
    lat = Latitude,
    color = "green", label = List
  )) |>
  # mean coordinates of World locations
  addCircleMarkers(
    lng = meanCoords[1, 2],
    lat = meanCoords[1, 1], color = "blue", label = "World mean",
  ) |>
  # mean coordinates of England locations
  addCircleMarkers(
    lng = meanCoords[2, 2],
    lat = meanCoords[2, 1], color = "blue", label = "England mean"
  ) |>
  # median coordinates of World locations
  addCircleMarkers(
    lng = medianCoords[1, 2],
    lat = medianCoords[1, 1], color = "red", label = "World median"
  ) |>
  # median coordinates of England locations
  addCircleMarkers(
    lng = medianCoords[2, 2],
    lat = medianCoords[2, 1], color = "red", label = "England median"
  )

worldMap
```

# 3a. Static plots

## Global view of places lived

```{r map worldwide plot}

# all locations
ggplot(dwellings, aes(x = Longitude, y = Latitude)) +
  geom_point(
    pch = 19,
    colour = dwellings$List,
    size = 2
  ) +
  coord_map(projection = "globular") +
  geom_vline(xintercept = -0.0076589) +
  geom_hline(yintercept = 51.4825766) +
  labs(
    title = "Coordinate plot 1: Worldwide locations",
    caption = "Note. Coordinates intercept at Greenwich, London"
  ) +
  theme(plot.caption = element_text(hjust = 0, face = "italic"))
```

## The spread of locations in England

```{r map uk plot}

# all England locations
ggplot(dwellings, aes(x = Longitude, y = Latitude)) +
  xlim(-2.5, 1.5) +
  ylim(51, 52.5) +
  geom_point(
    colour = dwellings$Latitude,
    size = 3
  ) +
  geom_text_repel(
    label = dwellings$List
  ) +
  coord_map(projection = "globular") +
  geom_vline(xintercept = -0.0076589) +
  geom_hline(yintercept = 51.4825766) +
  labs(
    title = "Coodinate plot 2: England",
    caption = "Note. Coordinates intercept at Greenwich, London."
  ) +
  theme(plot.caption = element_text(hjust = 0, face = "italic"))
```

You can see that most of the places lived in England have been between
$51^{\circ}$ and $52^{\circ}$ latitude, with a brief sojourn above the $52nd$
parallel (although I spent around five years living in Ipswich (16)); and
between $-2.5^{\circ}$ and $1.5^{\circ}$ longitude (quite a small area,
considering).

  - The numbers represent the order in which the places were lived and coincide
    with Table 1 above.

It is interesting that, although numerous; and despite the huge leap between
locations 2 and 5 (3 and 4 abroad, see main map above), the places lived are in
relatively tight clusters suggesting that I've enjoyed a local familiarity for a
while before needing/having to move on.

The clusters are divided across counties, as follows. 


# 3b. Breaking down into counties: 

## Essex and Suffolk

```{r map Essex plot}

# Essex and Suffolk
ggplot(dwellings, aes(x = Longitude, y = Latitude)) +
  xlim(-0.1, 1.3) +
  ylim(51.4, 52.2) +
  geom_point(
    colour = dwellings$List,
    size = 3
  ) +
  geom_text_repel(
    label = dwellings$List
  ) +
  coord_map(projection = "globular") +
  geom_vline(xintercept = -0.0076589) +
  geom_hline(yintercept = 51.4825766) +
  labs(
    title = "Coodinate plot 3: Essex and Suffolk",
    caption = "Note. Coordinates intercept at Greenwich, London"
  ) +
  theme(plot.caption = element_text(hjust = 0, face = "italic"))
```

## Surrey and Middlesex

```{r map Surrey/middx plot}

# Surrey and Middlesex
ggplot(dwellings, aes(x = Longitude, y = Latitude)) +
  xlim(-0.55, 0.001) +
  ylim(51.35, 51.49) +
  geom_point(
    colour = dwellings$List,
    size = 3
  ) +
  geom_text_repel(
    label = dwellings$List
  ) +
  coord_map(projection = "globular") +
  geom_vline(xintercept = -0.0076589) +
  geom_hline(yintercept = 51.4825766) +
  labs(
    title = "Coodinate plot 4: Surrey/Middlesex",
    caption = "Note. Coordinates intercept at Greenwich, London"
  ) +
  theme(plot.caption = element_text(hjust = 0, face = "italic"))
```

## Somerset

There's little point presenting a plot with two points on it, so Coordinate
plot 5 below shows the locations in Surrey/Middlesex and Somerset and the
expanse between.

```{r map Bath plot}

# Somerset
ggplot(dwellings, aes(x = Longitude, y = Latitude)) +
  xlim(-2.4, 0.01) +
  ylim(51.3, 51.5) +
  geom_point(
    colour = dwellings$List,
    size = 3
  ) +
  geom_text_repel(
    label = dwellings$List
  ) +
  coord_map(projection = "globular") +
  geom_vline(xintercept = -0.0076589) +
  geom_hline(yintercept = 51.4825766) +
  labs(
    title = "Coodinate plot 5: London to Somerset",
    caption = "Note. Coordinates intercept at Greenwich, London"
  ) +
  theme(plot.caption = element_text(hjust = 0, face = "italic"))
```

# 3c. Mean and Median locations

## Mean coordinates

I thought it would be interesting to know the mean and median locations of where
I have lived in the world; and in England.

Table 2 shows the mean coordinates of both the worldwide; and the England
locations from Table 1. In my mind, these give an indication of where I should
be living according to the mean-average coordinates of where I actually *have*
lived.

```{r mean coords table}

# mean coordinates
meanCoords |>
  kbl(caption = "<b>Table 2</b><br> <i>Average coordinates (mean)</i>") |>
  kable_classic_2(full_width = FALSE, position = "center")
```

```{r mean coords plot}

# mean coords plot
ggplot(meanCoords, aes(x = Longitude, y = Latitude)) +
  xlim(-1, 10) +
  ylim(49, 53) +
  geom_point(
    colour = "blue",
    size = 5
  ) +
  geom_text_repel(
    label = meanCoords$Place,
    nudge_x = 3, nudge_y = 0.5
  ) +
  coord_map(projection = "globular") +
  geom_vline(xintercept = -0.0076589) +
  geom_hline(yintercept = 51.4825766) +
  labs(
    title = "Coodinate plot 6: the average (mean) of all coordinates",
    caption = "Note. Coordinates intercept at Greenwich, London
            Cheshunt= England dwelling mean
            Génicourt-sur-Meuse= World dwelling mean"
  ) +
  theme(plot.caption = element_text(hjust = 0, face = "italic"))
```


I must admit, I quite fancy **Génicourt-sur-Meuse, France**. I've never been
there, but it sounds like it would suit me. Ironically, I worked in Cheshunt for
a fair while until I was made redundant, though I didn't miss the daily journey
from Harwich and back which became a drag after about 5 years of working for the
company there.


## Median coordinates

Table 3 shows the median coordinates of both the worldwide; and the England
locations from Table 1.

So, technically, these will be the absolute mid-point places to live according
to both options.

```{r median coords table}

# median coordinates
medianCoords |>
  kbl(caption = "<b>Table 3</b><br> <i> Average coordinates (median)</i>") |>
  kable_classic_2(full_width = FALSE, position = "center")
```

```{r med coords plot}

# median coordinates plot
ggplot(medianCoords, aes(x = Longitude, y = Latitude)) +
  xlim(-1, 2.5) +
  ylim(51, 52) +
  geom_point(
    colour = "red",
    size = 5
  ) +
  geom_text_repel(
    label = medianCoords$Place, 
    nudge_x = 1, nudge_y = 0.1
  ) +
  coord_map(projection = "globular") +
  geom_vline(xintercept = -0.0076589) +
  geom_hline(yintercept = 51.4825766) +
  geom_point(x = 0.176, y = 51.5655, pch = 8, size = 2) +
  labs(
    title = "Coodinate plot 7: the average (median) of all coordinates",
    caption = "Note. Coordinates intercept at Greenwich, London
            * = Rush Green, Dagenham
            Stapleford Abbotts= England dwelling median
            Upminster= World dwelling median"
  ) +
  theme(plot.caption = element_text(hjust = 0, face = "italic"))
```


<br>

> *Note.* The irony of the median coordinate plot is that both locations from
> the worldwide and England coordinates are a stones throw from where I was
> born! (Shown with an asterisk on the map above).

<br>

## Distances from the median locations to where I was born

So for fun, we'll use the `distm` function from the `geosphere` package to
calculate the distances between coordinates in metres (tutorial found on
[Statistics
Globe](https://statisticsglobe.com/geospatial-distance-between-two-points-in-r)
- actually a good site and recommended for many explanations of R-functions).

Once installed and loaded, I realised I needed another variable to hold the
\_Rush Green \_ coordinates. Then, I found out that `distm` needs the longitude
and latitude in that order, so had to reverse the two columns in my original
mean/median variables, then combined the two new variables into one, for each.

So the calculation returns like this for Stapleford Abbot: 


```{r distance between- setup}

# install.packages("geosphere") for calculating distance between coordinates

# load additional "geosphere" library
library(geosphere) 

# create rush_green coordinate variable
rush_green <- c(Longitude = 0.176, 
                Latitude = 51.5655, 
                Place = "Dagenham, Essex, England") 

# create new median coords variable and swap long. and lat. variables because that's 
# how `distm` likes it
median_coord_swap <- medianCoords |> 
  relocate(Longitude, .before = Latitude)

#  and combine row 2 with rush_to_stapleford variable
rush_to_stapleford <- median_coord_swap[2,] |> 
  rbind(rush_green)

# calculate distance between median places (Apply distm function)
dist_median <- distm(rush_to_stapleford[, 1:2], fun = distGeo)

# Print distance matrix
dist_median
```
...and this for Upminster...

```{r distance between- calc}

# new rush_to_upminster variable combined with row 1 of median swap variable
rush_to_upminster <- median_coord_swap[1,] |> 
  rbind(rush_green)

# calculate mean distance
dist_mean <- distm(rush_to_upminster[, 1:2], fun = distGeo)

# print matrix
dist_mean

```

## Summary

Remember the distances are in metres - I used [this
website](https://www.convertunits.com/from/meters/to/miles) (others are
available!) to convert the distances into miles (not sure if the site specifies
US miles as others have done, or my preferred UK miles, however...)

It's enough to give a guide to how close *Rush Green* is...

to the median of places in England:  

> __Stapleford Abbotts = 8.73Km (5.43 miles) away.__

to the median of worldwide locations:  

> __Upminster = 5.58Km (3.47 miles) away.__

This has been an interesting project - thanks for reading. I might take it
further one day and tell the story of each place and why I moved so many times.
I'm sure that the average moves during a lifetime is far fewer than what I have
achieved here. Does it mean I haven't settled anywhere? Or is it acting on
circumstances? A mixture of both? Why have I laid my hat in; and called so many
different locations, home?

(Social) psychology would describe these transitional moments of upheaval as
*ruptures* [(Zittoun, 2006,
p. 5)](https://www.google.co.uk/books/edition/Transitions/ivgnDwAAQBAJ?hl=en&gbpv=1&pg=PR1&printsec=frontcover),
since they can cause stress and disruption to 'normal' life (whatever *that*
is). Maybe that explains a few things!