Time & Place

Contents

• The Story: the why of it all
• The Timeline: the resulting timeline of how far we were from each other
• The Code: for the few interested, the code that produced the plot, and some comments
• Acknowledgements: it all started with a reticent fiancée and some duck-duck-going

---

The Story

We first “intentionally met” (bear with me) in the spring of 2013, and we’ve been together for nearly a decade now. While writing our story for our wedding we realized we might’ve had a chance encounter in 2003 in a random bus ride.

Taking this realization further, we noticed we hadn’t yet mapped all the places we’d been to before we first (intentionally) met, so we proceeded to do so, you know, in the unlikely event we’d unintentionally met more times. That’s when I had the idea of creating a timeline of places and the distances between us, which led to this little project.

First we just jotted down some dates and places in a spreadsheet and tried to manually order them, at which point I realized I’d already created an R script that orders two datasets based on a common key (in this case: date), so I threw the data into it and the rest is history.

---

The Timeline

Below is the visualization of major places we’ve spent at least a few nights - no more than one per month to keep things readable - and how far we were from each other in our lives before we met (yes, yes…) and started living and traveling together at an astounding 0 km from each other.

The Timeline

Back to Contents

---

The Code

For those interested, a brief look at how I put this together, starting with a simple csv.

Setup and Load

# Issue to Console:
# setwd('../GitHub/Timelines')
# rmarkdown::render('Time-And-Place.Rmd', output_file = 'Time-And-Place.html')

# cleanup environment
rm(list=ls())

# load libraries
install_packages <- function(package){
  
  newpackage <- package[!(package %in% installed.packages()[, "Package"])]
      
    if (length(newpackage)) {
        suppressMessages(install.packages(newpackage, dependencies = TRUE))
    }
    sapply(package, require, character.only = TRUE)
}


packages <- c("RJSONIO", "geosphere", "lubridate", "ggplot2", "tibble")
suppressPackageStartupMessages(install_packages(packages))

##   RJSONIO geosphere lubridate   ggplot2    tibble 
##      TRUE      TRUE      TRUE      TRUE      TRUE

# load simplified timeline
df <- read.csv("./data/timelines_simple.csv"
                , fileEncoding="UTF-8-BOM"
                , na.strings=c("NA","NULL","na","null","N/A","n/a","")
                , stringsAsFactors = FALSE)

# format date type
df$jdate <- as.Date(df$jdate, format="%m/%d/%Y")
df$mdate <- as.Date(df$mdate, format="%m/%d/%Y")

# peek
as_tibble(df)

## # A tibble: 38 x 8
##    jcity        jstate          jcountrycode jdate      mdate      mcity           mstate         mcountrycode
##    <chr>        <chr>           <chr>        <date>     <date>     <chr>           <chr>          <chr>       
##  1 Melbourne    Florida         US           1977-01-01 1977-01-01 São Paulo       São Paulo      BR          
##  2 Boulder      Colorado        US           1981-01-01 1990-06-01 Piracicaba      São Paulo      BR          
##  3 Fort Collins Colorado        US           1994-08-01 1990-07-01 São Paulo       São Paulo      BR          
##  4 Boulder      Colorado        US           1995-05-01 1991-12-01 Londrina        Paraná         BR          
##  5 Fort Collins Colorado        US           1995-08-01 1992-01-01 São Paulo       São Paulo      BR          
##  6 Sydney       New South Wales AU           1995-11-01 1992-12-01 Curitiba        Paraná         BR          
##  7 Fort Collins Colorado        US           1995-12-01 1993-01-01 São Paulo       São Paulo      BR          
##  8 Boulder      Colorado        US           1999-05-01 1994-01-01 Rio de Janeiro  Rio de Janeiro BR          
##  9 Kingston     Surrey County   JM           2003-06-01 1994-02-01 São Paulo       São Paulo      BR          
## 10 Boulder      Colorado        US           2003-07-01 1995-02-01 Puerto Quijarro Germán Busch   BO          
## # ... with 28 more rows

We simplified to a max of 1 location per month (of course there are many years without a single change) and all dates are the first of the month, that way I could make sure the y axis labels and graphics would be legible testing a single y value and checking the month spacing.

Wrangling

Here I create a new dataframe merging all dates in order, filling columns and rows as needed.

# get all dates, remove NAs and sort
all_dates <- c(df$jdate, df$mdate)
all_dates <- all_dates[!is.na(all_dates)]
all_dates <- sort(unique(all_dates))

# create a new dataframe
N <- length(all_dates)
dfm <- data.frame(
    'date' = all_dates
    , 'jcity' = rep(NA, N)
    , 'jstate' = rep(NA, N)
    , 'jcountrycode' = rep(NA, N)
    , 'mcity' = rep(NA, N)
    , 'mstate' = rep(NA, N)
    , 'mcountrycode' = rep(NA, N)
)

# fill cols
fill_col <- function(dfm, df, whosedate, col) {
    whosedate <- whosedate[!is.na(whosedate)]
    for (i in 1:nrow(dfm)) {
        dfm[i, colnames(dfm) == col] <- ifelse(dfm$date[i] %in% whosedate, 
            df[which(whosedate == dfm$date[i]), colnames(df) == col]
            , NA)
    }
    dfm
}

dfm <- fill_col(dfm, df, df$jdate, "jcity")
dfm <- fill_col(dfm, df, df$jdate, "jstate")
dfm <- fill_col(dfm, df, df$jdate, "jcountrycode")
dfm <- fill_col(dfm, df, df$mdate, "mcity")
dfm <- fill_col(dfm, df, df$mdate, "mstate")
dfm <- fill_col(dfm, df, df$mdate, "mcountrycode")

# fill all rows
fill_nas <- function(df, col) {
    for (i in 2:nrow(df)) {
        if (is.na(df[i, colnames(df) %in% col])) {
            df[i, colnames(df) %in% col] <- df[i-1, colnames(df) %in% col]
        }
    }
    df
}

dfm <- fill_nas(dfm, "jcity")
dfm <- fill_nas(dfm, "jstate")
dfm <- fill_nas(dfm, "jcountrycode")
dfm <- fill_nas(dfm, "mcity")
dfm <- fill_nas(dfm, "mstate")
dfm <- fill_nas(dfm, "mcountrycode")

# peek 
as_tibble(dfm)

## # A tibble: 71 x 7
##    date       jcity     jstate   jcountrycode mcity          mstate         mcountrycode
##    <date>     <chr>     <chr>    <chr>        <chr>          <chr>          <chr>       
##  1 1977-01-01 Melbourne Florida  US           São Paulo      São Paulo      BR          
##  2 1981-01-01 Boulder   Colorado US           São Paulo      São Paulo      BR          
##  3 1990-06-01 Boulder   Colorado US           Piracicaba     São Paulo      BR          
##  4 1990-07-01 Boulder   Colorado US           São Paulo      São Paulo      BR          
##  5 1991-12-01 Boulder   Colorado US           Londrina       Paraná         BR          
##  6 1992-01-01 Boulder   Colorado US           São Paulo      São Paulo      BR          
##  7 1992-12-01 Boulder   Colorado US           Curitiba       Paraná         BR          
##  8 1993-01-01 Boulder   Colorado US           São Paulo      São Paulo      BR          
##  9 1994-01-01 Boulder   Colorado US           Rio de Janeiro Rio de Janeiro BR          
## 10 1994-02-01 Boulder   Colorado US           São Paulo      São Paulo      BR          
## # ... with 61 more rows

Enrichment

Next I create a normalized table of locations avoid unnecessary API calls to openstreetmap where I get the latitude and longitudes of the cities in order to calculate the distances between them. I do this only once, given any location changes to the original CSV, so that in later code runs I can just pull up the normalized table from the data folder.

# only hit api if not exists (manually delete if changing input file)
if (!file.exists('./data/locs_df.csv')) {

    unique_locs <- unique(
        c(paste0(dfm$jcity, "|", dfm$jstate, "|", dfm$jcountrycode)
        , paste0(dfm$mcity, "|", dfm$mstate, "|", dfm$mcountrycode))
    )
    
    normalize_locs <- function(unique_locs) {
        N <- length(unique_locs)
        city <- state <- countrycode <- rep(NA, N)
        for (i in 1:N) {
            city[i] <- strsplit(unique_locs[i], "\\|")[[1]][1]
            state[i] <-  strsplit(unique_locs[i], "\\|")[[1]][2]
            countrycode[i] <-  strsplit(unique_locs[i], "\\|")[[1]][3]
        }
        df <- data.frame(
            'city' = city
            , 'state' = state
            , 'countrycode' = countrycode
        )
        df
    }
    
    locs_df <- normalize_locs(unique_locs)
    
    # instantiate lat long cols
    locs_df$lon <- locs_df$lat <- 0
    
    get_latlongs <- function(dfm) {
    
        for (i in 1:nrow(dfm)) {
    
            city_i <- gsub(' ', '%20', dfm$city[i])
            state_i <-  gsub(' ', '%20', dfm$state[i])
            countrycode_i <- dfm$countrycode[i]
    
            url <- paste0(
                "http://nominatim.openstreetmap.org/search?city=", city_i
                , "&state=", state_i
                , "&countrycodes=", countrycode_i
                , "&limit=4&format=json", sep="")
    
            data <- fromJSON(url)
    
            if (length(data) == 0) {
                print(paste0("Error requesting: ", url))
            }
            else if (is.vector(data)) {
                dfm$lon[i] <- as.numeric(data[[1]]$lon)
                dfm$lat[i] <- as.numeric(data[[1]]$lat)
            } 
            else {
                dfm$lon[i] <- as.numeric(data$lon)
                dfm$lon[i] <- as.numeric(data$lat)
            }
        }
        dfm
    }
    
    locs_df <- get_latlongs(locs_df)
    
    # save
    write.csv(locs_df, './data/locs_df.csv', row.names=FALSE)
}

# load locations
locs_df <- read.csv('./data/locs_df.csv')

# peek
as_tibble(locs_df)

## # A tibble: 38 x 5
##    city                    state           countrycode   lat     lon
##    <chr>                   <chr>           <chr>       <dbl>   <dbl>
##  1 Melbourne               Florida         US           28.1  -80.6 
##  2 Boulder                 Colorado        US           40.0 -105.  
##  3 Fort Collins            Colorado        US           40.6 -105.  
##  4 Sydney                  New South Wales AU          -33.9  151.  
##  5 Kingston                Surrey County   JM           18.0  -76.8 
##  6 Skye                    Highland        GB           57.4   -6.30
##  7 Wheat Ridge             Colorado        US           39.8 -105.  
##  8 Jemez                   New Mexico      US           35.6 -107.  
##  9 Kirribilli              New South Wales AU          -33.8  151.  
## 10 Saint Mary-of-the-Woods Indiana         US           39.5  -87.5 
## # ... with 28 more rows

Now I add the coordinates to the original dataframe, compute Haversine distances, and divide them in half for the plot. Distances are in kilometers, because science.

# enrich dfm with location data

# add j's lat, lon
dfm <- merge(
    dfm, locs_df
    , by.x=c('jcity', 'jstate', 'jcountrycode')
    , by.y = c('city', 'state', 'countrycode')
    , all = FALSE
)

dfm$jlat <- dfm$lat 
dfm$jlon <- dfm$lon 
dfm <- dfm[, !colnames(dfm) %in% c('lat','lon')]

# add m's lat, lon
dfm <- merge(dfm, locs_df, by.x=c('mcity', 'mstate', 'mcountrycode')
            , by.y = c('city', 'state', 'countrycode'), all = FALSE)
dfm$mlat <- dfm$lat 
dfm$mlon <- dfm$lon 
dfm <- dfm[, !colnames(dfm) %in% c('lat','lon')]

# sort by date
dfm <- dfm[order(dfm$date), ]

# Convert coords into distances

# calculate Haversine distance (in meters)
dfm$meters <- 0
for (i in 1:nrow(dfm)) {
    dfm$meters[i] <- distm(c(dfm$mlon[i], dfm$mlat[i])
        , c(dfm$jlon[i], dfm$jlat[i]), fun = distHaversine)
}

# calculate distance to midpoint (in km)
dfm$distance <- as.integer(round(dfm$meters / 2 / 1000))

# subset & reindex
dfm <- dfm[, c('date', 'jcity', 'mcity', 'jcountrycode', 'mcountrycode','distance')]
rownames(dfm) <- 1:nrow(dfm)

# peek
as_tibble(dfm)

## # A tibble: 71 x 6
##    date       jcity     mcity          jcountrycode mcountrycode distance
##    <date>     <chr>     <chr>          <chr>        <chr>           <int>
##  1 1977-01-01 Melbourne São Paulo      US           BR               3405
##  2 1981-01-01 Boulder   São Paulo      US           BR               4663
##  3 1990-06-01 Boulder   Piracicaba     US           BR               4594
##  4 1990-07-01 Boulder   São Paulo      US           BR               4663
##  5 1991-12-01 Boulder   Londrina       US           BR               4504
##  6 1992-01-01 Boulder   São Paulo      US           BR               4663
##  7 1992-12-01 Boulder   Curitiba       US           BR               4656
##  8 1993-01-01 Boulder   São Paulo      US           BR               4663
##  9 1994-01-01 Boulder   Rio de Janeiro US           BR               4753
## 10 1994-02-01 Boulder   São Paulo      US           BR               4663
## # ... with 61 more rows

Munging

Here I prepare some graphical parameters and merge into a tall-skinny dataframe at the month level (technically, first of the month).

# monthly and yearly frames
month_date_range <- seq(min(dfm$date), max(dfm$date), by='month')
month_format <- format(month_date_range, '%b')
month_df <- data.frame(month_date_range, month_format)

year_date_range <- c(seq(min(dfm$date) - years(1), max(dfm$date) + years(1), by='year'))
year_date_range <- as.Date(
    intersect(
        floor_date(year_date_range, unit="year")
        , ceiling_date(year_date_range, unit="year")
    ),  origin = "1970-01-01"
)
year_format <- format(year_date_range, '%Y')
year_df <- data.frame(year_date_range, year_format)

# create tall skinny date frame with all months
N <- length(month_date_range)
dfs <- data.frame(
    'date' = month_date_range
)

dfm$month <- as.integer(month(dfm$date))
dfm$year <- as.integer(year(dfm$date))

# Merge dfm into dfs frame for horizontal bars
mm <-  merge(dfs, dfm, by = c('date'), all = TRUE)
# reorder columns
mm <- mm[, c('date', 'month', 'year', 'jcity', 'jcountrycode', 'mcity', 'mcountrycode', 'distance')]

# fill out NAs
mm <- fill_nas(mm, "month")
mm <- fill_nas(mm, "year")
mm <- fill_nas(mm, "jcity")
mm <- fill_nas(mm, "jcountrycode")
mm <- fill_nas(mm, "mcity")
mm <- fill_nas(mm, "mcountrycode")
mm <- fill_nas(mm, "distance")

# peek 
as_tibble(mm)

## # A tibble: 439 x 8
##    date       month  year jcity     jcountrycode mcity     mcountrycode distance
##    <date>     <int> <int> <chr>     <chr>        <chr>     <chr>           <int>
##  1 1977-01-01     1  1977 Melbourne US           São Paulo BR               3405
##  2 1977-02-01     1  1977 Melbourne US           São Paulo BR               3405
##  3 1977-03-01     1  1977 Melbourne US           São Paulo BR               3405
##  4 1977-04-01     1  1977 Melbourne US           São Paulo BR               3405
##  5 1977-05-01     1  1977 Melbourne US           São Paulo BR               3405
##  6 1977-06-01     1  1977 Melbourne US           São Paulo BR               3405
##  7 1977-07-01     1  1977 Melbourne US           São Paulo BR               3405
##  8 1977-08-01     1  1977 Melbourne US           São Paulo BR               3405
##  9 1977-09-01     1  1977 Melbourne US           São Paulo BR               3405
## 10 1977-10-01     1  1977 Melbourne US           São Paulo BR               3405
## # ... with 429 more rows

Then I reshape into a single city and countrycode, doubling their values, distances, and adding directions for labeling and separating “who” those values refer to. I do this twice, once at the who-month level and once at the who-newplace level. The first is for the horizontal bars (distances) so I call it bars_df and the secon is for labeling changes in location so I call it labs_df.

# Reshape into single city, countrycode + distance + direction for labels for whom
# mm: tall with all dates for horiz bars
# dfm: only changing dates for city labels
reshape_df <- function(df) {

    out <- data.frame(
        'date' = rep(df$date, each = 2)
        , 'month' = rep(df$month, each = 2)
        , 'year' = rep(df$year, each = 2)
        , 'city' = c(rbind(df$jcity, df$mcity))
        , 'countrycode' = c(rbind(df$jcountrycode, df$mcountrycode))
        , 'distance' = c(rbind(df$distance, -df$distance))
        , 'direction' = as.integer(rep(c(1, -1), nrow(df)))
    )

    # factor country codes
    country_codes <- sort(unique(out$countrycode))
    out$countrycode <- factor(out$countrycode, levels = country_codes, ordered = TRUE)
    # add  'who' column
    who_levels <- c('Jenny', 'Marcelo')
    out$who <- rep(NA, nrow(out))
    out$who[out$direction == 1] <- who_levels[1]
    out$who[out$direction == -1] <- who_levels[2]
    out$who <- factor(out$who, levels = who_levels, ordered = TRUE)
    out <- out[, c('date', 'month', 'year', 'city', 'countrycode', 'distance', 'direction', 'who')]
    # return
    out
}

labs_df <- reshape_df(dfm)
bars_df <- reshape_df(mm)


# save
write.csv(labs_df, "./data/labs_df.csv", row.names = FALSE)
write.csv(bars_df, "./data/bars_df.csv", row.names = FALSE)

# peek
as_tibble(labs_df)

## # A tibble: 142 x 8
##    date       month  year city       countrycode distance direction who    
##    <date>     <int> <int> <chr>      <ord>          <int>     <int> <ord>  
##  1 1977-01-01     1  1977 Melbourne  US              3405         1 Jenny  
##  2 1977-01-01     1  1977 São Paulo  BR             -3405        -1 Marcelo
##  3 1981-01-01     1  1981 Boulder    US              4663         1 Jenny  
##  4 1981-01-01     1  1981 São Paulo  BR             -4663        -1 Marcelo
##  5 1990-06-01     6  1990 Boulder    US              4594         1 Jenny  
##  6 1990-06-01     6  1990 Piracicaba BR             -4594        -1 Marcelo
##  7 1990-07-01     7  1990 Boulder    US              4663         1 Jenny  
##  8 1990-07-01     7  1990 São Paulo  BR             -4663        -1 Marcelo
##  9 1991-12-01    12  1991 Boulder    US              4504         1 Jenny  
## 10 1991-12-01    12  1991 Londrina   BR             -4504        -1 Marcelo
## # ... with 132 more rows

Plot

First I reload the saved data and set up more graphical parameters, including the position of the text.

# reload data
load_data <- function(path) {
    read.csv(path
        , fileEncoding = 'latin1'
        , colClasses = c("Date", "integer", "integer", "character"
                        , "factor", "integer", "integer", "factor")
    )
}

bars_df <- load_data('./data/bars_df.csv')
labs_df <- load_data('./data/labs_df.csv')


# 10% max dist offset (used variously)
offset <- max(bars_df$distance)/10

# text position
bars_df$text_position <- (offset * 1.2 * bars_df$direction) + bars_df$distance
labs_df$text_position <- (offset * 1.2 * labs_df$direction) + labs_df$distance

# fix for close distances
fix_text_pos <- function(df) {
    df$text_position <- ifelse(df$direction == 1 & df$text_position < 2000, 2000 + offset, 
                        ifelse(df$direction == -1 & df$text_position > -2000, -2000 - offset
                        , df$text_position))
    df$text_position
}

bars_df$text_position <- fix_text_pos(bars_df)
labs_df$text_position <- fix_text_pos(labs_df)


# setup colors and levels
who_colors = c('blue3', 'darkred')
who_levels = c('Jenny', 'Marcelo')

# create arbitrary dataframe for meetings
meet_df <- data.frame(
    'date' = as.Date(c('2003-03-01', '2013-04-01'), format = '%Y-%m-%d')
    , 'distance' = as.integer(c(2500, 2500))
    , 'text' = c('Unintentional Meeting', 'Intentional Meeting')
)


# final peek
as_tibble(labs_df)

## # A tibble: 142 x 9
##    date       month  year city       countrycode distance direction who     text_position
##    <date>     <int> <int> <chr>      <fct>          <int>     <int> <fct>           <dbl>
##  1 1977-01-01     1  1977 Melbourne  US              3405         1 Jenny           4325.
##  2 1977-01-01     1  1977 São Paulo  BR             -3405        -1 Marcelo        -4325.
##  3 1981-01-01     1  1981 Boulder    US              4663         1 Jenny           5583.
##  4 1981-01-01     1  1981 São Paulo  BR             -4663        -1 Marcelo        -5583.
##  5 1990-06-01     6  1990 Boulder    US              4594         1 Jenny           5514.
##  6 1990-06-01     6  1990 Piracicaba BR             -4594        -1 Marcelo        -5514.
##  7 1990-07-01     7  1990 Boulder    US              4663         1 Jenny           5583.
##  8 1990-07-01     7  1990 São Paulo  BR             -4663        -1 Marcelo        -5583.
##  9 1991-12-01    12  1991 Boulder    US              4504         1 Jenny           5424.
## 10 1991-12-01    12  1991 Londrina   BR             -4504        -1 Marcelo        -5424.
## # ... with 132 more rows

Finally, I plot a very customized ggplot and save it as png.

# Plot
g <- ggplot(
        data = bars_df, aes(x = 0, y = date)
    ) +
    labs(color = "Who") +
    scale_color_manual(values = who_colors, labels = who_levels) +
    theme_classic()

# Plot: horizontal segment thick lines for all distances
g <- g + geom_segment(
    data = bars_df
        , aes(x = distance, xend = 0, yend = date)
        , color = 'azure2'
        , size = 8
)

# Plot: horizontal segment dotted lines for new cities
g <- g + geom_segment(
    data = labs_df
        , aes(x = distance, xend = 0, yend = date)
        , color = 'black'
        , size = 0.5
        , lty = 2
)

# Plot: scatter diamonds at timeline ends
g <- g +  geom_point(
        data = labs_df
        , aes(x = distance, color = who)
        , size = 5
        , pch = 18
)

# Plot: arbitrary scatter arrows at timeline ends
g <- g +  geom_point(
        data = meet_df
        , aes(x = 600, y = date)
        , color = 'maroon4'
        , size = 16
        , pch = -8592 # left arrow
)

# Plot: theme - less ink
g <- g + theme(
    legend.position = "top"
        , axis.title.x = element_blank()
        , axis.title.y = element_blank()
        , axis.line.x = element_blank()
        , axis.line.y = element_blank()
        , axis.text.x = element_blank()
        , axis.text.y = element_blank()
        , axis.ticks.x = element_blank()
        , axis.ticks.y = element_blank()
        , panel.background = element_rect(fill = "transparent")
        , plot.background = element_rect(fill = "transparent", color = NA)
        , legend.background = element_rect(fill = "transparent")
        , legend.title = element_text(size = 16)
        , legend.text = element_text(size = 14)
)

# Plot: vertical timeline
g <- g + geom_vline(
    xintercept = 0
        , color = "black"
        , size = 0.5
)

# Plot: year text
g <- g + geom_text(
    data = year_df
        , aes(x = -offset * 6/7, y = year_date_range, label = year_format, fontface = "bold")
        , check_overlap = TRUE
        , size = 4
        , vjust = 0
        , nudge_y = 5
        , color = "black"
)

# Plot: month text
g <- g + geom_text(
    data = month_df[seq(from = 2, to = nrow(month_df), by = 3), ]
        , aes(x = -offset * 1/4, y = month_date_range, label = month_format, fontface = "italic")
        , check_overlap = TRUE
        , size = 3
        , vjust = 0
        , nudge_y = 5
        , color = "black"
)

# Plot: city text
g <- g + geom_text(
   data = labs_df
       , aes(x = text_position, label = paste0(city, ', ', countrycode))
       , check_overlap = TRUE
       , size = 4
       # to override manual color, length must be 1 or df nrows
       , color = rep(who_colors, nrow(labs_df)/2)
)

# Plot: distance text
g <- g + geom_text(
   data = labs_df[labs_df$direction == 1, ]
       , aes(x = offset * 1/2, label = paste0(distance * 2, ' km'), fontface = "italic")
       , check_overlap = TRUE
       , size = 3.5
       , vjust = 0
       , nudge_y = 5
       # to override manual color, length must be 1 or df nrows
       , color = "black"
)

# Plot: meeting text
g <- g + geom_text(
   data = meet_df
       , aes(x = distance, label = text, fontface = "bold")
       , check_overlap = TRUE
       , size = 6
       , color = "maroon4"
)


# Save as png
ggsave(
    filename = "./img/timeline.png"
    , plot = last_plot()
    , device = "png"
    , width = 400
    , height = 2400
    , units = "mm"
    , limitsize = FALSE
    , dpi = 300
    , bg = "transparent"
)

# creates Rcode file with R code, set documentation = 1 to avoid text commentary
# library(knitr)
# options(knitr.purl.inline = TRUE)
# purl("Time-And-Place.Rmd", output = "Rcode.R", documentation = 1)

Back to Contents

---

Acknowledgements

I’m indebted to Ben Alex Keen for getting me started on this with his tutorial Creating A Timeline Graphic Using R and Ggplot2, and to my fiancée for digging up old dates and places she’d forgotten and putting them into a CSV.

Back to Contents

---