Lockdown and Air Pollution in Melbourne

background-image: url(https://cdn.pixabay.com/photo/2016/03/09/09/37/stars-1245902_960_720.jpg)
background-size: cover

## Lockdown and Air Pollution in Melbourne
## Team Gamma
### Authors: Chenjie Gong，Jimmy Effendy
<span style="color:white">
GitHub: [ <i class="fab  fa-github "></i> @chenjie](https://github.com/cgon0007)
 </span>

<span style="color:white">
GitHub: [ <i class="fab  fa-github "></i> @jimmy](https://github.com/j-effendy)
 </span>

### Date: 2020-11-05

---

# How has the Melbourne lockdown(s) affected air pollution around the city?

---
class: center,middle
# The Data

Microclimate Sensor Locations

**Provided by: City of Melbourne**
]

---
class:middle

background-image: url(http://res.co188.com/data/software/img640/5837307389885.jpg)

<img src="m.jpg" width="80%" />
#Assumption

Lockdown will always cause air pollution to be greatly reduced
---

# What is Air Pollution?
--

## PM2.5
--

## PM10
--

### SO2、NO2、O3

--
---
class:middle,center

# What is the standard? (AQI)
  
<img src="aqi.png" width="100%" />

---
### PM2.5 and PM10
PM stands for particulate matter (also called particle pollution)

The term for a mixture of solid particles and liquid droplets found in the air

PM10 have a diameter of 10 micrometres (0.01 mm) or smaller

PM2.5 have a diameter of 2.5 micrometres (0.0025 mm) or smaller

<img src="size.png" width="65%" />
---
class:middle,center

# Health Effects
.pull-left[

## PM2.5

#### Wheezing
#### Coughing
#### Chest tightness
#### Difficulty breathing.
]
--
.pull-right[

## PM10

#### High levels of PM10 can make you cough, your nose run and eyes sting
]
--
---
class:middle,center

# Sources of the Particles
.pull-left[
## PM2.5

#### Smoke from fires
#### Smoke from wood heaters
#### Car and truck exhausts
#### Industry
]
--
.pull-right[
## PM10

#### Dust from unsealed roads

#### Smoke from fires

#### Sea salt

#### Car and truck exhausts

#### Industry
]
--
---
class:middle,center
# Air Quality Category for PM2.5
<img src="2.5.png" width="85%" />

---
class:middle,center
# Air Quality Category for PM10
<img src="10.png" width="75%" />

---
### Sensor Location

There are five sensors

They are located along Grattan Street and Pelham Street

<div id="htmlwidget-b210b7a2a959c81c2859" style="width:504px;height:360px;" class="leaflet html-widget"></div>
<script type="application/json" data-for="htmlwidget-b210b7a2a959c81c2859">{"x":{"options":{"crs":{"crsClass":"L.CRS.EPSG3857","code":null,"proj4def":null,"projectedBounds":null,"options":{}}},"calls":[{"method":"addProviderTiles","args":["CartoDB.Voyager",null,null,{"errorTileUrl":"","noWrap":false,"detectRetina":false}]},{"method":"addCircleMarkers","args":[[-37.800793,-37.8023,-37.800629,-37.800575,-37.800524],[144.966492,144.960923,144.965052,144.964635,144.964122],10,null,null,{"interactive":true,"className":"","stroke":true,"color":"#03F","weight":5,"opacity":0.5,"fill":true,"fillColor":"#03F","fillOpacity":0.2},null,null,null,null,["Between 87 and 89 Grattan Street, Carlton","Corner of Pelham and Leicester Streets, Carlton","3/121 Grattan Street, Carlton","141 Grattan Street, Carlton","Corner of Swanston and Grattan Streets, Carlton"],{"interactive":false,"permanent":false,"direction":"auto","opacity":1,"offset":[0,0],"textsize":"10px","textOnly":false,"className":"","sticky":true},null]}],"limits":{"lat":[-37.8023,-37.800524],"lng":[144.960923,144.966492]}},"evals":[],"jsHooks":[]}</script>
---
### pm2.5 plot anglysis(the densityplot of pm2.5)

]
.panel[.panel-name[Learn]

Whether it is one-hour monitoring or one-day average monitoring,

the air quality in Melbourne is mostly good during the loackdown.

]
.panel[.panel-name[Code]

```r
d<-ggplot(pm25day_l, aes(x = value))+
  geom_density(aes(color=site_id))+
  xlim(1,33)+
  labs(title = "density plot of pm2.5_1day during the lockdowm")+
  xlab(label = "ug/m^3")+
  geom_vline(xintercept = 8)

d1<-ggplot(pm25_hourly, aes(x = value))+
  geom_density(aes(color=site_id))+
  xlim(1,33)+
  labs(title = "density plot of pm2.5_hourly during the lockdowm")+
  xlab(label = "ug/m^3")+
  geom_vline(xintercept = 27)
```

]
]

---
### pm2.5 plot anglysis(the linplot of pm2.5)

]
.panel[.panel-name[Learn]

PM2.5 did not decrease significantly during the blockade, but showed a small increase in winter.

The detection data of the five monitors are similar, and the accuracy of the monitor data is also verified.

]
.panel[.panel-name[Code]

```r
line<- pm25_hourly %>% filter(!month == "Jan") %>% 
  ggplot(aes(x = local_time,
             y = value),color=red) +
  geom_line(alpha = 0.6) +
  facet_wrap(~site_id)+
labs(x = "Time",
       y = "PM2.5 (μg/m3)",
       title = "Level of PM2.5 Particles in the different arc",
       subtitle = "Trends in 2020 (Excluding January)")
```

]
]

---
### pm2.5 plot anglysis(the calendar of pm2.5)
.panelset[
.panel[.panel-name[Plot]

]
.panel[.panel-name[Learn]
The air quality in Melbourne is mostly good during the loackdown.

In June and July, we found that the PM2.5 was relatively high, and this period was when the lockdown was very strict, and people were not allowed to go out into the City, so the lockdown did not have a big impact on air pollution.

]
.panel[.panel-name[Code]

```r
calendarPlot(pm25_day, 
             pollutant = "value", 
             breaks = c(0, 9.5, 25, 40, 177,500),  
             labels = c("good", "moderate", "poor", "Very poor","hazardous"),
             cols = c("green", "yellow", "orange",  "red","black"), 
             statistic="mean",
             year = 2020)
```

]
]

---

### PM10 Particle Concentration in Melbourne

]
.panel[.panel-name[Learn]

Huge spike of PM10 concentrations in January due to bushfires

Difficult to see trends of following months

]

```r
# Time-series plot for 2020
timeseries <- pm10_hourly %>%
  ggplot(aes(x = local_time,
             y = value)) +
  geom_line(color = "darkgrey") +
  geom_line(data = filter(pm10_hourly, !month == "Jan"),alpha = 0.6) +
  geom_vline(xintercept = as_datetime("2020-03-26"),
             color = "gold", size = 1) +
  geom_vline(xintercept = as_datetime("2020-07-08"),
             color = "orange", size = 1) +
  geom_vline(xintercept = as_datetime("2020-08-02"),
             color = "red", size = 1) +
  annotate("label", y = 750, x = as_datetime("2020-01-28"), label = "Bushfire", size = 4) +
  annotate("text", y = 900, x = as_datetime("2020-03-30"), label = "First restriction imposed \n 26 March 2020", size = 4, hjust = 0) +
  annotate("text", y = 750, x = as_datetime("2020-07-10"), label = "Stage 3 restriction \n 8 July 2020", size = 4, hjust = 0) +
  annotate("text", y = 400, x = as_datetime("2020-08-03"), label = "Stage 4 restriction \n 2 August 2020", size = 4, hjust = 0) +
  theme(aspect.ratio = 0.3) +
  labs(x = "Time",
       y = "PM10 (μg/m3)",
       title = "Timeseries plot",
       subtitle = "Trends in 2020") +
  scale_x_datetime(date_breaks = "1 month", date_labels = "%b")
```

]

---
### PM10 Particle Concentration in Melbourne

]
.panel[.panel-name[Learn]

A few instances of spikes in February due to bushfires

No significant change the first month after first lockdown

Some spikes in May

Air quality worsen after stage 3 restriction imposed

Air quality flatten after stage 4 restriction imposed

]
.panel[.panel-name[Code]

```r
timeseries2 <- pm10_hourly %>% filter(!month == "Jan") %>% 
  ggplot(aes(x = local_time,
             y = value)) +
  geom_line(alpha = 0.6) +
  geom_line(alpha = 0.6) +
  geom_vline(xintercept = as_datetime("2020-03-26"),
             color = "gold", size = 1) +
  geom_vline(xintercept = as_datetime("2020-07-08"),
             color = "orange", size = 1) +
  geom_vline(xintercept = as_datetime("2020-08-02"),
             color = "red", size = 1) +
  annotate("text", y = 170, x = as_datetime("2020-03-30"), label = "First restriction imposed \n 26 March 2020", size = 3.5, hjust = 0) +
  annotate("text", y = 165, x = as_datetime("2020-07-09"), label = "Stage 3 restriction \n 8 July 2020", size = 3.5, hjust = 0) +
  annotate("text", y = 110, x = as_datetime("2020-08-03"), label = "Stage 4 restriction \n 2 August 2020", size = 3.5, hjust = 0) +
  theme(aspect.ratio = 0.2) +
  labs(x = "Time",
       y = "PM10 (μg/m3)",
       title = "Timeseries plot",
       subtitle = "Trends in 2020 (Excluding January)") +
  scale_x_datetime(date_breaks = "1 month", date_labels = "%b")
```

]
]

---

### PM10 Particle Concentration in Melbourne

]
.panel[.panel-name[Learn]

Air quality in Melbourne are largely good

Some instances of very poor air quality in February, May, July, and October

No significant difference in the first two months after the lockdown in March

Higher median and variability in June and July

Significant decrease in PM10 concentration after Stage 4 restriction

]
.panel[.panel-name[Code]

```r
boxplot <- pm10_hourly %>% 
  filter(!month == "Jan") %>% 
  ggplot(aes(x = month,
             y = value,
             fill = lockdown_status)) +
  geom_boxplot() +
  geom_hline(yintercept = 120, color = "red") +
  geom_hline(yintercept = 80, color = "orange") +
  geom_hline(yintercept = 40, color = "gold") +
  annotate("text", y = 125, x = 7.5, label = "Very Poor", size = 3.5, alpha = 0.7) +
  annotate("text", y = 85, x = 8, label = "Poor", size = 3.5, alpha = 0.7) +
  annotate("text", y = 45, x = 8.5, label = "Moderate", size = 3.5, alpha = 0.7) +
  labs(y = "PM10 (μg/m3)",
       x = "Month",
       title = "Monthly Trend in 2020 (Excluding January)",
       subtitle = "Air quality reflected by the horizontal lines") +
  scale_y_continuous(breaks = c(20, 40, 50, 80, 100, 120, 150)) +
  scale_fill_brewer(name = "Restriction Level",
                    breaks = c("No Restriction", "First Restriction", 'Stage 3 Restriction', "Stage 4 Restriction")) + 
  theme(legend.position = "bottom")
```

]
]

---

### PM10 Particle Concentration in Melbourne

]
.panel[.panel-name[Learn]

#### Compared to no lockdown period

PM10 concentration during the first and stage 4 restrictions is lower

PM10 concentration during stage 3 restriction is higher

]
.panel[.panel-name[Code]

```r
boxplot2 <- pm10_hourly %>% 
  filter(!month == "Jan") %>% 
  ggplot(aes(x = fct_reorder(lockdown_status, -value),
             y = value)) +
  geom_boxplot(fill = "darkcyan",
               alpha = 0.6) +
  geom_hline(yintercept = 120, color = "red") +
  geom_hline(yintercept = 80, color = "orange") +
  geom_hline(yintercept = 40, color = "gold") +
  annotate("text", y = 131, x = 2.3, label = "Very Poor", size = 3.5, alpha = 0.8) +
  annotate("text", y = 87, x = 3.3, label = "Poor", size = 3.5, alpha = 0.8) +
  annotate("text", y = 50, x = 4.3, label = "Moderate", size = 3.5, alpha = 0.8) +
  labs(y = "PM10 (μg/m3)",
       x = "Restriction Level",
       title = "Level of PM10 particles for every restriction level",
       subtitle = "Air quality reflected by horizontal lines") +
  scale_y_continuous(breaks = c(20, 40, 50, 80, 100, 120, 150)) +
  coord_flip()
```

]
]

---

### Why are PM10 and PM2.5 Particulates More Concentrated in June and July?

Cold temperatures and still conditions may lead to increase in PM10 particulates

Lack of winds causes pollution from vehicles or wood-fired heaters to accumulate in the air (Environment Protection Authority, 2018)

The increase during this period is due to the increased demand for heating in winter and the effect of heat convection.

**Source: [ABC News](https://www.abc.net.au/news/2018-06-28/melbournes-fog-poor-air-quality-blamed-on-lack-of-wind/9920592)**

---

# Summary
The air quality in Melbourne is *good* in most days

PM10 and PM2.5 concentration during restriction periods are lower than the pre-lockdown period on average *except* during stage 3 restriction (June and July)

---
background-image: url(https://www.risk.net/sites/risk/files/styles/landscape_480_296/public/2019-10/Melbourne-central-business-district.jpg)
class: center, bottom, inverse
## You only live once!
### Go out :) (with social distancing) don't worry about air pollution, enjoy the fresh air!

---

# Thanks!
# Q&A