slides.qmd

---
title: "SiteSelection"
subtitle: "An open-source tool to identify complex areas for street space reallocation"
author: "Rosa Félix, Gabriel Valença, Filipe Moura, Ana Morais de Sá"
affiliation: "CERIS, Instituto Superior Técnico - University of Lisbon"
    # - name: "Rosa Félix"
    #   email: rosamfelix@tecnico.pt
    #   url: http://www.rosafelix.bike
    #   orcid: 0000-0002-5642-6006
    #   affiliation: "Instituto Superior Técnico - University of Lisbon"
    # - name: "Gabriel Valença"
    #   email: gabrielvalenca@tecnico.pt
    #   url: https://ushift.tecnico.ulisboa.pt/team-gabriel-valenca/
    #   orcid: 0000-0002-3684-3005
    # - name: "Filipe Moura"
    #   email: fmoura@tecnico.pt
    #   url: https://ushift.tecnico.ulisboa.pt/team-gabriel-valenca/
    #   orcid: 0000-0001-7749-8490
    # - name: "Ana Morais de Sá"
    #   email: ana.morais.sa@tecnico.pt
    #   url: https://ushift.tecnico.ulisboa.pt/team-gabriel-valenca/
    #   orcid: 0000-0003-0591-9369
    #   # affiliation: "Instituto Superior Técnico - University of Lisbon"
execute-dir: project
format:
  revealjs: 
    theme: default
    slide-number: true
    chalkboard: 
      buttons: false
    preview-links: auto
    logo: img/ist-ceris.png
    # css: styles.css
    footer: '22^nd^ GET meeting, Alcobaça 2025'
# editor: visual
bibliography: paper/NECTAR/mybibfile.bib
repo-url: https://github.com/U-Shift/EITcourse
# cache: true # erro com os mapviews
---

## Complex areas

-   Space is limited, and there is **competition** 🚶‍♂️🚗🚲🚌🎍

-   Growing need for **more sustainable**, efficient, and equitable **space distribution**

-   Challenge of **reallocating** street space, dynamically over time

. . .

-   Complexity of identifying areas where intervention makes sense [@Valenca24]

## Combination of several types of data

A process to select the **cell locations** that meet a set o criteria for a given city or neighbourhood where the street space is more disputed.

::: incremental
-   Population density (census)
-   POIs and activities
-   Public Transit Frequency (GTFS for bus and tram)
-   Road network centrality measures
    -   betweenness, closeness, and degree
:::

<!-- ![](images/clipboard-1767020115.png){.absolute bottom="300" right="50" width="600"} -->

## Problem {background-color="none"}

**Using *point and click* software to fetch and process data...**

This is fine for a single case study 😀

![](images/clipboard-1066373156.png){fig-align="center" width="600"}

. . .

But what it you want to replicate the **exactly** same process for **400** other locations? 🤯


::: footer
:::

## `SiteSelection`

> SiteSelection aims to identify **complex areas** suitable for street space reallocation.

. . .

**GIS tools:** methods to process existing geo-data and classify areas based on key metrics like road network centrality, population density, and public transport frequency.

. . .

**Data Sources**: open data such as *OpenStreetMap* road network and POIs, Census data, GTFS for public transport, administrative areas

. . .

**Open-source tool**: an `R` package 📦

<!-- ## Processes -->

<!-- ```{r} -->
<!-- targets::tar_visnetwork() -->
<!-- ``` -->

## Processes

```{r}
targets::tar_visnetwork(targets_only = TRUE)
```

## Map

```{r}
targets::tar_load(grid_all)
mapview::mapview(grid_all, zcol="score", alpha = 0.7) # ranked
```

::: {style="font-size: 80%;"}
Returns a ranked list of complex areas based on their need for space reallocation (0 - no complexity, 6 - very complex)
:::

## Options

#### Thresholds for candidate selection

``` {.r code-line-numbers="1-3|4|5|6"}
degree_min = mean           # mean or median? default: mean
betweeness_range = 0.40     # percentile to exclude (upper and lower) default: 0.25
closeness_range = 0.25      # percentile to exclude (upper and lower) default: 0.25
population_min = median     # mean or median? default: mean
entropy_min = 0.35          # value to exclude (lower) default: 0.5
freq_bus = c(4, 10, 20)     # frequency of bus stops to define level of service 
```

![](images/clipboard-1269564533.png){fig-align="center"}

::: {.callout-warning appearance="simple"}
For different contexts, we can have different cut thresholds... *Under assessment*
:::

## Options

### Grid types

-   Squares
-   Hexagons
-   Universal [h3 hexagon grid](https://h3geo.org/) - useful to include other variables <!-- -   All can **vary resolution** -->

::: {layout-ncol="3"}
![](images/clipboard-772809911.png){width="280"}

![](images/clipboard-3306234436.png)

![](images/clipboard-2647626512.png){width="280"}
:::

## Options

### **Work with non-administrative boundary**

-   Set another polygon limit to the analysis
-   [create a `.geojson` file](https://geojson.io/) with the area of interest

```{r}
GEOJSON = TRUE          # use a different limit? in inputdata/*.geojson
GEOJSON_input = "map1"  # name of the file if GEOJSON = TRUE. default: "map1"
```

::: {layout-ncol="2"}
```{r}
mapview::mapview("inputdata/map1.geojson")
# mapview::mapview("outputdata/map1/grid_all.gpkg", zcol="score")
```

![](images/clipboard-3611128328.png)
:::

::: footer
:::

## Options

Keep track of your analysis and export results

```{r}
analysis = TRUE # export input parameters and results to a xls file? default: FALSE
```

![](images/clipboard-2210141522.png)

## Live demo

-   Modify a network parameter

-   Modify the grid type or resolution

-   Other case study


## Conclusions

::: incremental
-   Repeated processes and analysis, human error proof 🛡️
-   Super fast 🚀
    -   It would normally take days to run everything 😵
    -   \~1min for a city
    -   for 200 cities em Portugal \< 45 min
:::

## Conclusions

::: incremental
-   Change one parameter and re-run **only what is needed**
-   Scalable and expandable, for other custom conditions and functions
-   Replicable for other locations 🌎🌍🌏
    -   We included scripts to prepare data (census, GTFS)
-   Open source 📡
:::

## Future developments and applications

-   Each function can standalone in the package

    -   `get_census()`
    -   `get_osm()`, `clean_osm()`
    -   `get_centrality()`
    -   `make_grid()`
    -   `get_transit()`

. . .

-   And be used for other purposes! out of the *Streets4All* scope

## Thank you

#### More: [u-shift.github.io/SiteSelection](https://u-shift.github.io/SiteSelection)

![contributions are welcome!](images/clipboard-4136130744.png){fig-align="center"}

::: {align="center"}
**Rosa Félix -** [rosamfelix\@tecnico.pt](maito:)
:::

## About

This work is part of [Streets4All Project](https://streets4all.tecnico.ulisboa.pt/), developed at the University of Lisbon and at the University of Coimbra, and funded by Fundação para a Ciência e Tecnologia (*PTDC/ECI-TRA/3120/2021*).

[![](images/clipboard-3372171257.png){fig-align="center"}](https://streets4all.tecnico.ulisboa.pt/)

::: {style="font-size: 40%;"}
This work is part of the research activity carried out at Civil Engineering Research and Innovation for Sustainability (CERIS) and has been funded by Fundação para a Ciência e a Tecnologia (FCT), Portugal in the framework of project UIDB/04625/2020, project PTDC/ECI-TRA/3120/2021, and project 2022,07909,CEECIND/CP1713/CT0017.
:::

</br> The concept is based on: