Spatial Analysis

class: inverse, bottom, right

<h1 align="center"> Spatial Analysis Using R </h1>

<h3> Montevideo - June 2019 </h3>

<a href="https://orlando-sabogal.github.io/">Orlando Sabogal</a> 
 -
 <a href="https://twitter.com/AntonioSabogal">@AntonioSabogal</a> 
 -
 <a href="https://github.com/orlando-sabogal">Github</a>

???
This is the first silde
Introduce myself

---

#What is special about spatial data?

- The title of this slide was taken from [An old but good article](https://escholarship.org/uc/item/3ph5k0d4) by Luc Anselin.

--
- Tobler's (1979) first law of geography. Have you heard of the second law?
- What *near* and *distant* mean in a particular context?
- Data-driven vs model-driven

--
- What about independece and correlation among observations?

--
- Computers and geography are in the same team. See [Geography and computers: Past, present, and future](https://onlinelibrary.wiley.com/doi/10.1111/gec3.12403) by Arribas-Bel and Reades 
- [Goegraphic Data Science article](https://onlinelibrary.wiley.com/doi/full/10.1111/gean.12194) by Singleton and Arriba-bel
- There is much that geography can learn about data science.

--
- Data is **NOT ENOUGH** to extract knowledge.

---
class: inverse, center, middle
#Visualization

---
class: inverse, center, middle
background-image: url(Minard.png)
background-size: contain

???
Source:

---
class: inverse, center, middle
background-image: url(bike_ggplot_james-cheshire-768x508.png)
background-size: contain

???
Source: https://www.r-graph-gallery.com/270-james-cheshire-dataart/
see: http://oobrien.com/wp-content/uploads/2012/09/mappinglondon_soc.pdf
---
class: inverse, center, middle
background-image: url(journey_to_work_web_james_cheshire-768x898.gif)
background-size: contain

???
Source: https://www.r-graph-gallery.com/270-james-cheshire-dataart/

---
#Statistical models

--
- Areal data: Observed variable on zones follows a random process. Spatial Autucorrelation. Modifable areal unit problem MAUP.

--
- Point data:Location as a random process (crimes). Differnce between spatial distribution and completely random spatial distribution. Usually associated with hotspots and Kernel Denstiy Estimation KDE.

--
- Continous data: *Geostatistics.* The variables is continously distributed across all the territory but there are limited discrete-fixed observations. Semivariogram, interpolation, Kriging.

---
#Programming

- Programming? Writting code? or just *scripting*?

- Always use google. Even the most experimented programmers are all day in stackoverflow or github figuring out how to make things work.

- I see programming as a superpower.

- It can certainly foster your work.

- Reproducible research, government accountability, citizen science.

- **Empowerment**

- R is a (powerful) tool. Be aware of other tools: *python*, *javascript*, *SQL.*

---
#Content

This course is about using data with geographic properties. 
Hopefully, **we will code a lot.**

.pull-left[

###You will learn

- Some general concepts. 
- Key Libraries
- To make reproducible and reusable maps
- That *the sky is the limit*
- Hopefully, **to learn R**.

]

.pull-right[

###You will not learn

- Geographic Statistics and sound methods.
- *raster* data.
- To write efficient R code.*
- To handle big data.
- To solve every possible problem.

]

.footnote[[*] This is up to you.]

---
#libraries

--
There are many libraries. In this course we are going to try to cover the basics apsects of:

--
sf 
--
sf + *The Tidyverse* 
--
sf + *The Tidyverse* + notebooks 
--
tmap 
--
leaflet

---
class: inverse, center, middle

# sf: simple features

---
#Spatial Operations

--
- Read and Write Data

```r
st_read()
st_write()
```

--
- It works as a tibble with and additional *geometry* variable. You can use almost the whole **Tidyverse**

```r
SF_Object %>% ...functions()...

filter()
arrange()
group_by()
summarise()
...
```

---
###Some Useful Functions

```r
plot(st_geometry())
st_join()
st_union()
st_difference()
st_buffer()
...
```

--
###CRS: Coordinate Reference System
- *geographic* (distance is not in meters)
- *projected* (distance is in meters)

```r
st_crs()
st_is_longlat()

st_set_crs()
st_transform()
```
--
Go to the [sf Tutorial!](https://orlando-sabogal.github.io/SpatialAnalysis-MontevideoWorkshop2019/Notebooks/WarmUp.nb.html)

---
class: inverse, center, middle

#tmap
##(I love it)

---
#A grammar for plots

**tmap** works in a similar way to **ggplot2**. 
Is simple to use. Think about having an empty canvas where you add elements by using **+ tm_SomeFuncion.**

```r
tm_shape() + tm_polygons() +
  tm_shape() + tm_dots() + 
  tm_compass() + tm_scale_bar()
```

###Some capabilities:
- Auxiliar maps (insets)
- Interactiviy, connection to Leaflet
- Full customization of legends
- Faceted maps

Go to the [tmap Tutorial!](https://orlando-sabogal.github.io/SpatialAnalysis-MontevideoWorkshop2019/Notebooks/tmap-tutorial.nb.html)

---
class: inverse, center, middle

#leaflet

---
#Leaflet and Leaflet for R

Originallyl, Leaflet is a javascript [library](https://leafletjs.com/) "...designed with simplicity, performance and usability in mind."

Fortunately, in the [html widgets for R](http://www.htmlwidgets.org/) there is a [Leaflet for R](https://rstudio.github.io/leaflet/) that let you use most of the capabilities of Leaflet with your regular R code. To create maps with leaflet you can use *data.frame* objects, *sf* objects, *raster* objects and *json* files.

Leaflet for R is methodologically easy to use and you can get farther by exploring the arguments of the functions. A nice feature of Leaflet is that is integrated with shiny.

```r
leaflet() %>% addProviderTiles() %>% 
  addPolygons() %>% 
  addCircles() %>% 
  ...
```

---
#Some examples

- [The propensity to cycling tool PCT](http://www.pct.bike/m/?r=avon)
- [PCT website](https://www.pct.bike/)
- [PCT Github repositories!](https://github.com/npct) 
- What about a similar tool for Uruguay?

- [Reimagining NYC Neighborhoods with NewerHoods](https://data-clinic.shinyapps.io/newerhoods/) [Repo](https://github.com/tsdataclinic/newerhoods) 
- [Isolines](https://byollin.shinyapps.io/Isolines/). [Repo](https://github.com/byollin/Isolines)
- [Water Quality](https://udwq.shinyapps.io/lakeprofiledashboard/)
- [SpatialEpiApp](https://paulamoraga.shinyapps.io/spatialepiapp/)
- Is actually an [R package](https://paula-moraga.github.io/blog/2018/01/04/2018-01-04_spatialepiapp/) too.

--
Go to the [leaflet Tutorial!](https://orlando-sabogal.github.io/SpatialAnalysis-MontevideoWorkshop2019/Notebooks/InteractiveMaps.nb.html)

---
class: inverse, center, middle

#The sky is the limit!

---

###There are many **free** and **open** resources that you can use:

- [Geocomputation with R](https://geocompr.robinlovelace.net/) 
- [Spation-Temporal Statistics With R](https://spacetimewithr.org/)
- [Spatial Data Science](https://keen-swartz-3146c4.netlify.com/)
- [Predictive Soil Mapping With R](https://soilmapper.org/)
- [Spatial Data Science](https://rspatial.org/index.html)
- [Intro to Gis and Spatial Analysis](https://mgimond.github.io/Spatial/index.html)
- [Workshop on R and Movement Ecology](https://www.danaseidel.com/MovEco-R-Workshop/)
- [Geoinformatics with R](http://rgeo.tbuckl.com/)

---
class: center, middle

# Thanks!

Slides created via the R package [**xaringan**](https://github.com/yihui/xaringan).

The chakra comes from [remark.js](https://remarkjs.com), [**knitr**](http://yihui.name/knitr), and [R Markdown](https://rmarkdown.rstudio.com).