Spatial_join.knit

# Spatial Join
## Claudia Gutiérrez-Arellano
### 30 April 2026 | The University of Manchester 
---
# Which areas in Greater Manchester are best served by the tram network?

<p class="image-caption">
Source: TfGM Metrolink Network Map

</p>

</div>

---

Tram stops (points) and boroughs (polygons)

<img src="data:image/png;base64,#index_files/figure-html/unnamed-chunk-1-1.png" width="80%" style="display: block; margin: auto;" />
<p class="image-caption">
GM Metrolink Network, Source: data.gov.uk; Local Authority Districts (Dec 2022), Source: ONS Geography
---

# Spatial join

- Combines attributes from one spatial dataset with another

- **Key feature:**  Based on **location**, not shared data

---

# Spatial join

- Combines attributes from one spatial dataset with another

- **Key feature:**  Based on **location**, not shared data

Which boroughs are not served by the tram network?
  
  Which borough has the most tram stops?
  
  Which tram stops are located within each borough?
]

.pull-right[
<div style="text-align: center;">
<img src="images/gmtram.png" width="90%">
<p class="image-caption">
Source: Open Street Map, ONS Combined Authorities, TfGM
</div>
]

---
# Spatial join process

## Inputs - target & source (same Coordinate Reference System)
- Tram stops as **points**
- Boroughs as **polygons**

## Spatial rule
- For each tram stop, identify the borough it falls **within**

## Output
- 'Stops in borough' as **points**  
  Each tram stop inherits the attributes of a polygon
  
---
# Spatial join  in R
.pull-left[
Libraries

``` r
library(sf)         # simple features
library(tidyverse)  # data management
```
Function: `st_join`

Our Inputs
- target: `stops`
- source: `boroughs`

Our Rule: `st_within` (but there are many more!)

Our Output: `stops_in_borough`]
.pull-right[
Let's run!

``` r
stops = st_read("geodata/Metrolink_Stops_Functional.json", quiet = TRUE)
boroughs = st_read("geodata/GM_lad.gpkg", quiet = TRUE)

# check they have the same reference
st_crs(boroughs) == st_crs(stops) 
#[1] TRUE

stops_in_borough = st_join(
  stops,          
  boroughs,       
  join = st_within 
)
```
]

---

## Checking the output

### Before: 
.pull-left[

`stops`
<table class="table" style="font-size: 16px; margin-left: auto; margin-right: auto;">
 <thead>
  <tr>
   <th style="text-align:left;"> name </th>
   <th style="text-align:left;"> stationCode </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Heaton Park </td>
   <td style="text-align:left;"> HPK </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Bowker Vale </td>
   <td style="text-align:left;"> BKV </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Crumpsall </td>
   <td style="text-align:left;"> CRP </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Bury </td>
   <td style="text-align:left;"> BRY </td>
  </tr>
</tbody>
</table>

]

`boroughs`
<table class="table" style="font-size: 16px; margin-left: auto; margin-right: auto;">
 <thead>
  <tr>
   <th style="text-align:left;"> LAD22CD </th>
   <th style="text-align:left;"> LAD22NM </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> E06000007 </td>
   <td style="text-align:left;"> Warrington </td>
  </tr>
  <tr>
   <td style="text-align:left;"> E06000008 </td>
   <td style="text-align:left;"> Blackburn with Darwen </td>
  </tr>
  <tr>
   <td style="text-align:left;"> E06000049 </td>
   <td style="text-align:left;"> Cheshire East </td>
  </tr>
  <tr>
   <td style="text-align:left;"> E07000037 </td>
   <td style="text-align:left;"> High Peak </td>
  </tr>
</tbody>
</table>
]
<br>

<span style="font-size: 16px;">
*Use `glimpse(input_name)` for a quick check
<span style="font-size: 16px;">
---

## Checking the output
.pull-left[
### After:

`stops_in_borough`
<table class="table" style="font-size: 16px; margin-left: auto; margin-right: auto;">
 <thead>
  <tr>
   <th style="text-align:left;"> name </th>
   <th style="text-align:left;"> stationCode </th>
   <th style="text-align:left;"> LAD22CD </th>
   <th style="text-align:left;"> LAD22NM </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Heaton Park </td>
   <td style="text-align:left;"> HPK </td>
   <td style="text-align:left;"> E08000002 </td>
   <td style="text-align:left;"> Bury </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Bowker Vale </td>
   <td style="text-align:left;"> BKV </td>
   <td style="text-align:left;"> E08000003 </td>
   <td style="text-align:left;"> Manchester </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Crumpsall </td>
   <td style="text-align:left;"> CRP </td>
   <td style="text-align:left;"> E08000003 </td>
   <td style="text-align:left;"> Manchester </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Bury </td>
   <td style="text-align:left;"> BRY </td>
   <td style="text-align:left;"> E08000002 </td>
   <td style="text-align:left;"> Bury </td>
  </tr>
</tbody>
</table>
<br>

<span style="font-size: 16px;">
*Use `glimpse(output_name)` for a quick check
<span style="font-size: 16px;">
]

<img src="data:image/png;base64,#index_files/figure-html/unnamed-chunk-6-1.png" width="130%" style="display: block; margin: auto;" />
]

---

## Using the joined data to answer questions

``` r
borough_counts = stops_in_borough %>%
  st_drop_geometry() %>%
  count(LAD22NM, name = "n_stops")%>%
  arrange(-n_stops)
```

.pull-left[
<table class="table" style="font-size: 12px; margin-left: auto; margin-right: auto;">
 <thead>
  <tr>
   <th style="text-align:left;"> LAD22NM </th>
   <th style="text-align:right;"> n_stops </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Manchester </td>
   <td style="text-align:right;"> 42 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Trafford </td>
   <td style="text-align:right;"> 18 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Oldham </td>
   <td style="text-align:right;"> 10 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Salford </td>
   <td style="text-align:right;"> 10 </td>
  </tr>
</tbody>
</table>
]

<table class="table" style="font-size: 12px; margin-left: auto; margin-right: auto;">
 <thead>
  <tr>
   <th style="text-align:left;"> LAD22NM </th>
   <th style="text-align:right;"> n_stops </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Tameside </td>
   <td style="text-align:right;"> 7 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Bury </td>
   <td style="text-align:right;"> 6 </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Rochdale </td>
   <td style="text-align:right;"> 6 </td>
  </tr>
</tbody>
</table>
]
]

Which boroughs are not served?

``` r
not_served = boroughs %>%
  st_drop_geometry() %>%
  anti_join(stops_in_borough, by = "LAD22NM") %>%
  select(LAD22NM)
```
<div class="three-col"> <div> 
<table class="table" style="font-size: 12px; margin-left: auto; margin-right: auto;">
 <thead>
  <tr>
   <th style="text-align:left;"> LAD22NM </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Warrington </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Blackburn with Darwen </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Cheshire East </td>
  </tr>
  <tr>
   <td style="text-align:left;"> High Peak </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Chorley </td>
  </tr>
</tbody>
</table>
</div> <div>

<table class="table" style="font-size: 12px; margin-left: auto; margin-right: auto;">
 <thead>
  <tr>
   <th style="text-align:left;"> LAD22NM </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Rossendale </td>
  </tr>
  <tr>
   <td style="text-align:left;"> West Lancashire </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Bolton </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Stockport </td>
  </tr>
</tbody>
</table>
</div> <div> 
<table class="table" style="font-size: 12px; margin-left: auto; margin-right: auto;">
 <thead>
  <tr>
   <th style="text-align:left;"> LAD22NM </th>
  </tr>
 </thead>
<tbody>
  <tr>
   <td style="text-align:left;"> Wigan </td>
  </tr>
  <tr>
   <td style="text-align:left;"> St. Helens </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Calderdale </td>
  </tr>
  <tr>
   <td style="text-align:left;"> Kirklees </td>
  </tr>
</tbody>
</table>
</div> </div>

<span style="font-size: 16px;">
`anti_join` returns all rows from `boroughs` without a match
</span>

]

---

.pull-left[
# Spatial join in Python 
- Same inputs
- Same spatial rule: within
- Same type of output

<br>

]

``` python
import geopandas as gpd

stops_in_borough = gpd.sjoin(
    stops,
    boroughs,
    how="inner",
    predicate="within"
)

borough_counts = (
    stops_in_borough
    .groupby("LAD22NM")
    .size()
    .reset_index(name="n_stops")
)

not_served = (
    boroughs[["LAD22NM"]]
    .merge(stops_in_borough[["LAD22NM"]], on="LAD22NM", how="left", indicator=True)
    .query('_merge == "left_only"')
    [["LAD22NM"]]
)
```
]

---

.pull-left[
# Spatial join in Python 
- Same inputs
- Same spatial rule: within
- Same type of output

<br>

]

``` python
import geopandas as gpd

stops_in_borough = gpd.sjoin(
    stops,
    boroughs,
    how="inner",
    predicate="within"
)

borough_counts = (
    stops_in_borough
    .groupby("LAD22NM")
    .size()
    .reset_index(name="n_stops")
)

not_served = (
    boroughs[["LAD22NM"]]
    .merge(stops_in_borough[["LAD22NM"]], on="LAD22NM", how="left", indicator=True)
    .query('_merge == "left_only"')
    [["LAD22NM"]]
)
```
]

---
class: middle
## Other spatial operations

| Spatial relationship | R (sf) | Python (GeoPandas) | Example |
|---------------------|--------|--------------------|---------|
| point within polygon | `join = st_within` | `predicate="within"` | Which schools are located within flood risk zones? |
| features overlap | `join = st_intersects` | `predicate="intersects"` | Which green spaces overlap proposed development zones? |
| boundaries touch | `join = st_touches` | `predicate="touches"` | Which protected areas border urban areas? |
| nearest feature* | `st_nearest_feature` | `sjoin_nearest()` | Which hospital is closest to a sports facility? |

<span style="font-size: 20px;">
Explore: 
<a href="https://r-spatial.github.io/sf/reference/st_join.html" target="_blank">sf st_join</a> |
<a href="https://geopandas.org/en/stable/docs/reference/api/geopandas.GeoDataFrame.sjoin.html" target="_blank">geopandas sjoin</a>
<span style="font-size: 20px;">

---

# Summary

- A spatial join links datasets using **location**

- It follows a **process**: inputs > spatial rule > output  
   
- We used a spatial join to identify the boroughs that are/are not served by the tram network

- The same process applies in **R** and **Python**

---
class: middle
## A question for you

If a borough contains a tram stop, does that mean people across the whole borough have good access to the network?

---
class: middle
## Not necessarily

- Presence of stops `\(\neq\)` accessibility

- Other analyses (wait to hear about `st_nearest_feature`, `st_distance` and `st_buffer`!)

---
<br><br><br><br>

<h1 style="text-align: center;">
Questions?
</h1>

.pull-left[ 
<span style="font-size: 20px;">
Access R and Python code, data and slides 
<a href="https://github.com/Claudia-Gutierrez/spatial-join-presentation.git" target="_blank">here</a> 
<span style="font-size: 16px;">
]