05_update.Rmd

---
title: "Impact of COVID-19 on refugees and migrants, Update 5"
author: "Mixed Migration Centre, 30 June 2020"
date: ''
output: github_document
 
---


```{r setup, include=FALSE}
knitr::opts_chunk$set(echo = TRUE)
library(tinytex)
```
\  

**This is the fifth update on the situation for refugees and migrants on mixed migration routes around the world in light of the COVID-19 pandemic, based on data collected by the [Mixed Migration Centre](http://www.mixedmigration.org/).  As MMC moves to a new phase in its data collection, this update looks at changes over time on the themes covered in the Global Updates since April: COVID-19 awareness, knowledge and risk perception, access to healthcare, assistance needs and the impact on refugees’ and migrants’ lives and migration journeys. It also includes data on Afghan returnees. New Global Updates will be available soon, and for more detailed, thematic and response-oriented COVID-19 snapshots from each of the MMC regional offices, see [here](http://www.mixedmigration.org/resource-type/covid-19/).**


## Key Messages
• Knowledge of COVID-19 remains stable, and concern is high among refugees and migrants, although there has been a decrease in reports of fear of transmission. At the same time, the proportion of respondents not taking measures to protect themselves from the disease is falling

• The proportion of respondents reporting barriers to healthcare is falling, and especially in Latin America, a greater share are reporting that they can access healthcare

• There are suggestions that some aspects of day-to-day life may be normalizing (although the limitations on the data must be taken into consideration). Fewer are reporting a reduced availability of basic goods, fewer are reporting loss of income (especially in North Africa and West Africa), and fewer are saying that loss of income is impacting on their ability to afford basic goods. Stress and anxiety, however, have been increasingly reported over time

• Inability to continue the journey is increasingly reported, except in North Africa, where the proportion has dropped considerably over time. However, the proportion reporting that the crisis has not impacted on their journey has increased, as has the proportion reporting that the crisis has not impacted on their plans. Fewer people are reporting that they have decided to stay where they are for the time being

• The proportion reporting needs has stayed consistently very high. The need for cash is most frequently reported, and has grown. While assistance received remains much lower than what is needed, the proportion receiving cash has grown. The perceived need for information has fallen.

## Respondents

4,124 respondents were interviewed between 6 April and 8 June 2020, with 302 in Asia, 161 in East Africa, 646 in Latin America, 1,886 in North Africa, and 1,129 in West Africa:

```{r table, echo=FALSE}
table_demographics <- data.frame(Region=c("Asia", "", "", "East Africa", "", "Latin America", "", "North Africa", "", "West Africa", "", "", "Overall"), Country=c("India", "Indonesia", "Malaysia", "Kenya", "Somaliland", "Colombia", "Peru", "Libya", "Tunisia", "Burkina Faso", "Mali", "Niger", ""), n=c(112, 117, 73, 59, 102, 496, 150, 935, 951, 327, 422, 380, "4,124"), `Percent women` = c(38, 39, 41, 46, 35, 74, 60, 30, 36, 43, 18, 29, 38), `Mean age`=c(33, 29, 27, 33, 31, 34, 33, 30, 28, 28, 27, 31, 30))
library(knitr)
kable(table_demographics, align = 'c', col.names = c("Region", "Country", "Number of respondents", "Percent women", "Mean age"))


```

In Latin America, all respondents were Venezuelans, while in Asia, 76% of respondents were Afghans (15% were from Myanmar, and 9% from Bangladesh). Data collection in Malaysia only began in period 2. Respondents in East Africa, North Africa and West Africa were from a wide range of African countries.

For the purpose of this update looking at trends over time (see methodology), the respondents were split in the following three time periods:

```{r table2, echo=FALSE}
table_demo <- data.frame(Period=c(1, 2, 3), From=c("06/04", "06/05", "21/05"), To=c("05/05", "20/05", "08/06"), Asia=c(60, 145, 97), `East Africa` = c(37, 57, 67), `North Africa`=c(988, 596, 302), `West Africa` = c(586, 326, 217), `Latin America`=c(377, 60, 209))
library(knitr)
kable(table_demo, align = 'c', col.names = c("Period", "From", "To", "Asia", "East Africa", "North Africa", "West Africa", "Latin America"))

```


## Methodology

A summary of the methodology can be found [here](http://www.mixedmigration.org/4mi/4mi_faq/). For this update, the whole dataset was split into three time periods. It is important to note that the respondents were not the same over the three time periods (i.e. we did not interview the same respondents several times as in a panel or longitudinal study). Respondents were however recruited in the same locations, from the same target population and using the same method. All figures are rounded to the nearest whole number. Figures for countries where the number of interviews is less than or around 100 should be interpreted with caution, especially in East Africa, and period 1 in Asia, and period 2 in Latin America. Unless specified, the number of observations for all analyses and visualisations corresponds to those presented in the above table. Note that for most items of the questionnaire, respondents can select several answer options. 88 interviews were discarded from analyses due to questionnaire incompleteness or data quality issues. 

## Awareness, knowledge and risk perception

Overall, the number of respondents who agree or strongly agree that they are worried about catching coronavirus has remained remarkably stable over time, at around 89%. However, we can see a decreasing trend in especially West Africa (from 90% at Time 2 to 83% at Time 3).

The number of respondents who agree or strongly agree that they are worried about transmitting coronavirus has tended to decrease over time (from 70% at Time 1 to 67% at Time 3), see Figure 1. This is the case in all regions, except East Africa and Asia, where the number of interviews is low. The sharpest decrease was in North Africa, from 68% to 55%.

**_Figure 1:  Percentage of respondents who agree or strongly agree that they are worried about transmitting coronavirus, over time_**

```{r 0, include=FALSE}
library(tidyverse)
load("rda/05_cleaned_data_20200609.rda")
#CREATE PERIODS####
#split
data  <- data %>% 
  mutate(Period = ifelse(Date >= '2020-04-06' & Date <= '2020-05-05', 1,
                ifelse(Date >= '2020-05-06' & Date <= '2020-05-20', 2,
                3)))
mean(data$Period)
dim(data) #208 cols

#count
data %>% group_by(Period, Region) %>% tally()
data %>% group_by(Period, Region) %>% tally() %>% 
  mutate(Percent=n/sum(n)*100) %>%
  ggplot(aes(x=as.factor(Period), y=Percent, fill=Region))+
  geom_col(width = 0.5)


#add n by period
data <- data %>% 
  group_by(Period) %>% 
  mutate(N_period = length(Period))
dim(data) #209 cols
#add n by period and region
data <- data %>% 
  group_by(Period, Region) %>% 
  mutate(N_period_region = length(Region))
dim(data) #210 cols
#add n by period and country
data <- data %>% 
  group_by(Period, Q13_1c) %>% 
  mutate(N_period_country = length(Q13_1c))
dim(data) #211 cols
```

```{r 1, echo=FALSE, fig.align='left', fig.height=8, fig.width=8, warning=FALSE}
data$c4 <- factor(data$c4, levels = c("Strongly disagree", "Disagree", "Neither agree nor disagree", "Agree", "Strongly agree", "Refused"))



#region
c4 <- data %>% 
  mutate(c4 = recode(c4, "Strongly agree" = "Agree")) %>% 
  group_by(Period, Region, N_period_region, c4) %>% 
  tally() %>% 
  mutate(Percent=round(n/N_period_region*100, digits = 1)) %>%
  filter(c4=="Agree") %>% 
  ungroup() %>% 
  add_row(Period=1, Region="Overall", N_period_region=2048, c4="Agree", n=1435, Percent=70.1) %>% 
  add_row(Period=2, Region="Overall", N_period_region=1184, c4="Agree", n=820, Percent=69.3) %>%
  add_row(Period=3, Region="Overall", N_period_region=892, c4="Agree", n=602, Percent=67.5) 

c4$Region <- factor(c4$Region,levels = c("Asia", "East Africa", "Latin America", "North Africa", "West Africa", "Overall"))

fig2 <- c4 %>% 
  ggplot(aes(x=as.factor(Period), y=Percent, color=Region)) +
  geom_line(aes(group=Region, linetype=Region), size=2)+
  scale_y_continuous(limits = c(0, 100), breaks = seq(0, 100, by = 10))+
  scale_linetype_manual(values=c("solid", "solid", "solid", "solid", "solid", "dashed"))+
  scale_color_manual(values=c("#CC79A7", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#000000"))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))
fig2

```
\  
\

As observed in previous updates, the number of respondents who reportedly know how to protect themselves is high, and has remained stable over time, with a slight increase from the second (83%) to third time period (87%). Given the efforts to provide information about prevention since the pandemic was declared, an increase in knowledge could have been expected, even though respondents already showed a high level of knowledge when data collection started.

Likewise, the overall proportion of respondents reporting not doing anything to protect themselves against coronavirus has remained low over time, but with interesting differences between regions. In West Africa, there seems to be a clear decreasing trend (from 20% at Time 1 to 13% at Time 3), whereas in East Africa, the figure has increased over the three time periods (from 8% to 27%). 

In all regions but one, there has been a fall in the percentage of respondents staying at home to protect themselves from catching coronavirus (East Africa: from 38% to 30%; Latin America: from 88% to 74%; North Africa: from 54% to 36%; West Africa: from 8% to 4%). While lockdown measures have somewhat eased in some locations, the fall may also be attributable to people needing to leave their home more, e.g. to work. In Asia, by contrast, the figure increased from 80% in the first period to 87% in the last period. This is likely linked to the fact that later time periods include data from Malaysia.  


## Access to healthcare and prevention

When looking at the data over time (Figure 2), we see that the proportion of people who believe they could not access healthcare, overall, has slightly decreased (from 31% at Time 1 to 29% at Time 3), except in Asia (from 38% to 21%) and Latin America (from 40% to 25%). Note that the inclusion of data from Malaysia from period 2 only is likely to have influenced this more dramatic drop, as respondents from Malaysia have reported better access to healthcare.

**_Figure 2:  Percentage of respondents who believe they could access healthcare, over time_**

```{r 2, echo=FALSE, fig.align='left', fig.height=8, fig.width=13, warning=FALSE}
c17  <- data %>%
  group_by(Period, Region, N_period_region, c17) %>% 
  tally() %>% 
  mutate(Percent=round(n/N_period_region*100, digits = 1)) %>% 
  filter(c17=="Yes" | c17=="No" | c17=="Don´t know") %>% 
  ungroup() %>% 
  add_row(Period=1, Region="Overall", N_period_region=2048, c17="Don´t know", n=569, Percent=27.8) %>% 
  add_row(Period=2, Region="Overall", N_period_region=1184, c17="Don´t know", n=405, Percent=34.2) %>%
  add_row(Period=3, Region="Overall", N_period_region=892, c17="Don´t know", n=246, Percent=27.6) %>% 
  add_row(Period=1, Region="Overall", N_period_region=2048, c17="No", n=627, Percent=30.6) %>% 
  add_row(Period=2, Region="Overall", N_period_region=1184, c17="No", n=362, Percent=30.6) %>%
  add_row(Period=3, Region="Overall", N_period_region=892, c17="No", n=259, Percent=29) %>% 
  add_row(Period=1, Region="Overall", N_period_region=2048, c17="Yes", n=792, Percent=38.7) %>% 
  add_row(Period=2, Region="Overall", N_period_region=1184, c17="Yes", n=401, Percent=33.9) %>%
  add_row(Period=3, Region="Overall", N_period_region=892, c17="Yes", n=376, Percent=42.2)

c17$Region <- factor(c17$Region,levels = c("Asia", "East Africa", "Latin America", "North Africa", "West Africa", "Overall"))
c17$c17 <- factor(c17$c17,levels = c("Yes", "No", "Don´t know"))

fig5 <- c17 %>%
  ggplot(aes(x=as.factor(Period), y=Percent, color=Region)) +
  geom_line(aes(group=Region, linetype=Region), size=2)+
  scale_y_continuous(limits = c(0, 100), breaks = seq(0, 100, by = 10))+
  scale_linetype_manual(values=c("solid", "solid", "solid", "solid", "solid", "dashed"))+
  scale_color_manual(values=c("#CC79A7", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#000000"))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))+
  facet_wrap(~c17)+
  theme(strip.text = element_text(size=14))
fig5

```

\  
\

Reflecting this change, Latin America saw a large increase in respondents believing they would be able to access healthcare from 37% (n=22) during the second period to 61% (n=127) in the last period, which could at least partially be explained by information campaigns. The proportion of respondents who simply don’t know whether they would be able to access healthcare has also tended to remain roughly stable over time, with a peak at 34% during the second time period.

Since the beginning of data collection, the main barriers to accessing healthcare cited by respondents are a lack of money, followed by not knowing where to go, and discrimination against foreigners. This does not change when the data are analysed over time, as can be seen in Figure 3. However, all barriers are less frequently reported over time, which aligns with the slight increase in respondents who identified no barrier (from 8% to 11%).


**_Figure 3: What are the barriers to accessing healthcare?_**

```{r 3, echo=FALSE, fig.align='left', fig.height=8, fig.width=9.5, warning=FALSE}
library(RColorBrewer)

fig6 <- data %>% ungroup() %>% #need to ungroup first, which is odd
  select(208, 209, 115:128) %>% #this excludes nones
  pivot_longer(cols = 3:16, names_to="Options", values_to="Answer") %>%
  filter(!is.na(Answer)) %>% 
  group_by(Period, N_period, Answer) %>%
  tally() %>% 
  mutate(Percent=round(n/N_period*100)) %>%
  filter(Answer != "Refused") %>% 
  filter(Answer != "Don't know") %>%
  filter(Answer != "Other (specify)") %>%
  mutate(Answer = recode(Answer, "I don't know where to go for healthcare"="I don't know where to go", "The advice for testing and treating coronavirus is unclear here"="The advice is unclear","I don't have the money to pay for health services"="I don't have the money","I don't have the right or the legal documentation to access health services here"="I don't have the right", "I am afraid of being reported to authorities, or arrest, or deportation"="Afraid of being reported", "Discrimination against foreigners limits access to services"="Discrimination against foreigners","General insecurity and conflict prevent me from accessing healthcare"="General insecurity and conflict","There are no health services here"="There are no health services","Services are overwhelmed and access is difficult for everyone"="Services are overwhelmed",)) %>%
  ggplot(aes(x=as.factor(Period), y=Percent, color=Answer)) +
  geom_point(aes(group=Answer))+
  geom_line(aes(group=Answer), size=2)+
  scale_y_continuous(breaks = seq(0, 100, by = 5))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  scale_color_brewer(palette = "Paired")+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))
fig6

```

\  
\


## Impact on refugees’ and migrants’ lives

The main impacts of COVID-19 on refugees’ and migrants’ lives have been reduced access to work, more stress, and reduced availability of basic goods, and this has remained very stable over time.

Racism and xenophobia have tended to slightly decrease overall (from 20% in the first period to 17% in the last period), which seems to be mainly due to North Africa (from 26% to 21%), but was stable in all other regions, including in Latin America, where a 10% decrease during the second period is, arguably, only due to the low number of interviews at this time (n=60).

More clearly, the impact on availability of basic goods has decreased from the first period to the last period, although it remains high, particularly in Asia (Asia: from 68% to 60%; East Africa: from 68% to 43%; North Africa: from 58% to 56%; West Africa: from 28% to 21%). In contrast, respondents in Latin America have more frequently reported an impact on the availability of basic goods. The overall trends (see Figure 4) fit with the patterns of assistance needed vs assistance received, below.

**_Figure 4:  Percentage of respondents citing reduced availability of basic goods as an impact, over time_**

```{r 4, echo=FALSE, fig.align='left', fig.height=8, fig.width=8, warning=FALSE}
c20_bis <- data %>% ungroup() %>% #need to ungroup first, which is odd
  select(208, 205, 210, 171:178) %>% #this includes nones
  pivot_longer(cols = 4:11, names_to="Options", values_to="Answer") %>% 
  filter(!is.na(Answer)) %>% 
  group_by(Period, Region, N_period_region, Answer) %>% 
  tally() %>% 
  mutate(Percent=round(n/N_period_region*100, digits = 1)) %>% 
  filter(Answer=="Reduced availability of basic goods") %>% 
  ungroup() %>% 
  add_row(Period=1, Region="Overall", N_period_region=2048, Answer="Reduced availability of basic goods", n=1039, Percent=50.7) %>% 
  add_row(Period=2, Region="Overall", N_period_region=1184, Answer="Reduced availability of basic goods", n=555, Percent=46.9) %>%
  add_row(Period=3, Region="Overall", N_period_region=892, Answer="Reduced availability of basic goods", n=441, Percent=49.4) 

c20_bis$Region <- factor(c20_bis$Region,levels = c("Asia", "East Africa", "Latin America", "North Africa", "West Africa", "Overall"))

fig13  <- c20_bis %>% 
  ggplot(aes(x=as.factor(Period), y=Percent, color=Region)) +
  geom_line(aes(group=Region, linetype=Region), size=2)+
  scale_y_continuous(limits = c(0, 100), breaks = seq(0, 100, by = 10))+
  scale_linetype_manual(values=c("solid", "solid", "solid", "solid", "solid", "dashed"))+
  scale_color_manual(values=c("#CC79A7", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#000000"))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))
fig13

```
\  
\

Another potentially encouraging trend is that the proportion of respondents who lost income due to coronavirus restrictions has slightly decreased over time (from 60% in the first period to 57% in the last period). This includes the two regions with more interviews, North Africa (from 57% to 49%) and West Africa (from 47% to 40%). Latin America, where so far respondents more often reported a loss of income compared to other regions, has remained stable (from 90% to 91%). In Asia, loss of income has been more frequently reported, again perhaps because of the addition of data from Malaysia (from 42% to 49%), see Figure 5.


**_Figure 5: Percentage of respondents who lost income due to coronavirus restrictions, over time_**

```{r 5, echo=FALSE, fig.align='left', fig.height=8, fig.width=8, warning=FALSE}
c21 <- data %>%
  group_by(Period, Region, N_period_region, c21) %>% 
  tally() %>% 
  mutate(Percent=round(n/N_period_region*100, digits = 1)) %>%  
  filter(c21=="Yes") %>% 
  ungroup() %>% 
  add_row(Period=1, Region="Overall", N_period_region=2048, c21="Yes", n=1226, Percent=59.9) %>% 
  add_row(Period=2, Region="Overall", N_period_region=1184, c21="Yes", n=637, Percent=53.8) %>%
  add_row(Period=3, Region="Overall", N_period_region=892, c21="Yes", n=505, Percent=56.6)

c21$Region <- factor(c21$Region,levels = c("Asia", "East Africa", "Latin America", "North Africa", "West Africa", "Overall"))

fig14 <- c21 %>%
  ggplot(aes(x=as.factor(Period), y=Percent, color=Region)) +
  geom_line(aes(group=Region, linetype=Region), size=2)+
  scale_y_continuous(limits = c(0, 100), breaks = seq(0, 100, by = 10))+
  scale_linetype_manual(values=c("solid", "solid", "solid", "solid", "solid", "dashed"))+
  scale_color_manual(values=c("#CC79A7", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#000000"))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))
fig14

```
\  
\

The share of respondents who report losing income in each region has remained stable overall. In contrast, the inability to afford basic goods has been reported far less frequently in Asia (from 84% in the first period to 43% in the last period), whereas it tended to be stable or slightly increase in other regions, see left panel in Figure 6. Additionally, the inability to continue the journey has tended to increase in most regions, except in North Africa (from 41% in the first period to 26% in the last period, see middle panel in Figure 6). Finally, the inability to pay remittances has remained stable or has tended to increase (in North Africa), but has decreased in the regions with fewer interviews (Asia and East Africa), see right panel in Figure 6.

**_Figure 6: Impacts of loss of income by region, over time_**

```{r 6, echo=FALSE, fig.align='left', fig.height=8, fig.width=13, warning=FALSE}
c22 <- data %>% ungroup() %>% #need to ungroup first, which is odd
  filter(c21 == "Yes") %>% #discard those who don't need help
  select(208, 205, 181:188) %>% 
  group_by(Period, Region) %>% 
  mutate(n_region_p = length(Period)) %>% ungroup() %>%
  pivot_longer(cols = 3:10, names_to="Options", values_to="Answer") %>%
  filter(!is.na(Answer)) %>%
  group_by(Period, Region, Answer, n_region_p) %>% tally() %>%
  mutate(Percent=round(n/n_region_p*100, digits = 1)) %>% 
  filter(Answer=="I am unable to afford basic goods" | Answer=="I am unable to pay remittances" | Answer=="I am unable to continue my journey") %>%
  ungroup() %>% 
  add_row(Period=1, Region="Overall", n_region_p=1226, Answer="I am unable to afford basic goods", n=819, Percent=66.8) %>% 
  add_row(Period=2, Region="Overall", n_region_p=637, Answer="I am unable to afford basic goods", n=389, Percent=61.1) %>%
  add_row(Period=3, Region="Overall", n_region_p=505, Answer="I am unable to afford basic goods", n=357, Percent=70.7) %>% 
  add_row(Period=1, Region="Overall", n_region_p=1226, Answer="I am unable to pay remittances", n=474, Percent=38.7) %>% 
  add_row(Period=2, Region="Overall", n_region_p=637, Answer="I am unable to pay remittances", n=216, Percent=33.9) %>%
  add_row(Period=3, Region="Overall", n_region_p=505, Answer="I am unable to pay remittances", n=199, Percent=39.4) %>% 
  add_row(Period=1, Region="Overall", n_region_p=1226, Answer="I am unable to continue my journey", n=419, Percent=34.2) %>% 
  add_row(Period=2, Region="Overall", n_region_p=637, Answer="I am unable to continue my journey", n=231, Percent=36.3) %>%
  add_row(Period=3, Region="Overall", n_region_p=505, Answer="I am unable to continue my journey", n=130, Percent=25.7)

c22$Region <- factor(c22$Region,levels = c("Asia", "East Africa", "Latin America", "North Africa", "West Africa", "Overall"))

fig16 <- c22 %>% 
  ggplot(aes(x=as.factor(Period), y=Percent, color=Region)) +
  geom_line(aes(group=Region, linetype=Region), size=2)+
  scale_y_continuous(limits = c(0, 100), breaks = seq(0, 100, by = 10))+
  scale_linetype_manual(values=c("solid", "solid", "solid", "solid", "solid", "dashed"))+
  scale_color_manual(values=c("#CC79A7", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#000000"))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))+
  facet_wrap(~Answer)+
  theme(strip.text = element_text(size=14))
fig16

```

\


## Assistance needs

The need for extra assistance has remained very high and has slightly increased over time (from 86% to 88%), except in North Africa, where on the contrary it has tended to slightly decrease (from 85% to 80%). 

\
**_Figure 7: Types of assistance needed, over time_**

```{r 7, echo=FALSE, fig.align='left', fig.height=8, fig.width=9, warning=FALSE}
fig8  <- data %>% ungroup() %>% #need to ungroup first, which is odd
  filter(c26 == "Yes") %>%
  select(208, 157:169) %>% 
  group_by(Period) %>% 
  mutate(n_period = length(Period)) %>% ungroup() %>%
  pivot_longer(cols = 2:14, names_to="Options", values_to="Answer") %>%
  filter(!is.na(Answer)) %>% 
  mutate(Answer = recode(Answer, "Access to health services"="Health services", "Distribution of sanitary items (sanitizer/ mask/ gloves/ etc)"="Sanitizer, masks, gloves", "Cash to pay for health services"="Cash for health services", "Documentation to access health services"="Documentation", 'Information about the virus: symptoms/ what to do if I have symptoms/ how to protect myself'="Information", "Other basic needs: food, water, shelter"="Food, water, shelter", "Access to work and livelihoods"="Access to work", "Psychological assistance"="Psychological support", "Other (specify)"="Other")) %>%
  filter(Answer != "Refused") %>% 
  filter(Answer != "Other") %>% 
  group_by(Period, Answer, n_period) %>% tally() %>%
  mutate(Percent=round(n/n_period*100, digits = 1)) %>% 
  ggplot(aes(x=as.factor(Period), y=Percent, color=Answer)) +
  geom_line(aes(group=Answer), size=2)+
  scale_y_continuous(breaks = seq(0, 100, by = 10))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  scale_color_brewer(palette = "Paired")+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))
fig8

```
\  
\

As shown in Figure 7, cash is the most reported need. The need has grown between the first period and the last period in all regions, except in Asia, where it decreased (from 71% to 63%), see Figure 8.

Sanitary/PPE items such as sanitizer, masks, and gloves, which were overall the third most-cited need, is also the only need that has clearly tended to decrease over time (from 42% in the first period to 30% in the last period), see Figure 7, which is in line with the slightly decreasing proportion of respondents concerned about transmitting the virus.

**_Figure 8:  Percentage of respondents who cite cash as extra assistance needed, over time_**

```{r 8, echo=FALSE, fig.align='left', fig.height=8, fig.width=8, warning=FALSE}
c27 <- data %>% ungroup() %>% #need to ungroup first, which is odd
  filter(c26 == "Yes") %>% #discard those who don't need help
  select(208, 205, 157:169) %>% 
  group_by(Period, Region) %>% 
  mutate(n_region_p = length(Period)) %>% ungroup() %>%
  pivot_longer(cols = 3:15, names_to="Options", values_to="Answer") %>%
  filter(!is.na(Answer)) %>%
  group_by(Period, Region, Answer, n_region_p) %>% tally() %>% 
  mutate(Percent=round(n/n_region_p*100, digits = 1)) %>% 
  filter(Answer=="Cash") %>%
  ungroup() %>% 
  add_row(Period=1, Region="Overall", n_region_p=1770, Answer="Cash", n=1327, Percent=75) %>% 
  add_row(Period=2, Region="Overall", n_region_p=1012, Answer="Cash", n=806, Percent=79.6) %>%
  add_row(Period=3, Region="Overall", n_region_p=782, Answer="Cash", n=608, Percent=77.5)

c27$Region <- factor(c27$Region,levels = c("Asia", "East Africa", "Latin America", "North Africa", "West Africa", "Overall"))

fig9  <- c27 %>% 
  ggplot(aes(x=as.factor(Period), y=Percent, color=Region)) +
  geom_line(aes(group=Region, linetype=Region), size=2)+
  scale_y_continuous(limits = c(0, 100), breaks = seq(0, 100, by = 10))+
  scale_linetype_manual(values=c("solid", "solid", "solid", "solid", "solid", "dashed"))+
  scale_color_manual(values=c("#CC79A7", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#000000"))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))
fig9

```
\  
\

As highlighted in previous updates, although more than 85% of all participants consistently state that they need extra assistance, less than 25% actually received such extra assistance since beginning of data collection. However, an encouraging trend is emerging when analysing the data over time. Although assistance received was stable in the first two periods (20% and 21%, respectively), it has clearly increased in the last period, with 33% (n=294) of respondents reporting to have benefited from it. Furthermore, assistance received has increased in all regions (Figure 9), except in East Africa (from 19% to 3%), although the number of interviews is low in this region.

**_Figure 9: Percentage of respondents who received extra assistance, over time_**

```{r 9, echo=FALSE, fig.align='left', fig.height=8, fig.width=8, warning=FALSE}
c23 <- data %>%
  group_by(Period, Region, N_period_region, c23) %>% 
  tally() %>% 
  mutate(Percent=round(n/N_period_region*100, digits = 1)) %>% 
  filter(c23=="Yes") %>% 
  ungroup() %>% 
  add_row(Period=1, Region="Overall", N_period_region=2048, c23="Yes", n=415, Percent=20.3) %>% 
  add_row(Period=2, Region="Overall", N_period_region=1184, c23="Yes", n=251, Percent=21.2) %>%
  add_row(Period=3, Region="Overall", N_period_region=892, c23="Yes", n=294, Percent=33)

c23$Region <- factor(c23$Region,levels = c("Asia", "East Africa", "Latin America", "North Africa", "West Africa", "Overall"))

fig10  <- c23 %>% 
  ggplot(aes(x=as.factor(Period), y=Percent, color=Region)) +
  geom_line(aes(group=Region, linetype=Region), size=2)+
  scale_y_continuous(limits = c(0, 100), breaks = seq(0, 100, by = 10))+
  scale_linetype_manual(values=c("solid", "solid", "solid", "solid", "solid", "dashed"))+
  scale_color_manual(values=c("#CC79A7", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#000000"))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))
fig10

```
\  
\

Over time, the most frequently cited types of assistance received have remained the same: basic needs, sanitary items, and cash. The only exception is information about the virus, which has decreased over time (from 20% in the first period to 10% in the last period). This fits well with the observed trend of fewer respondents citing a need for information, see Figure 7 above. Given that cash assistance is the most cited need, it is encouraging to see a slight increase in the proportion of respondents who received cash assistance (from 23% to 25%). When looking at trends in each region, extra assistance received in terms of basic needs has decreased in Asia (from 62% in the first period to 40% in the last period, although the number of interviews is low). In contrast, the proportion of respondents saying they received sanitary items have increased in West Africa (from 42% in the first period to 61% in the last period), although fewer respondents in West Africa actually report this as a need (from 49% in the first period to 43% in the last period).

## Impact on migration journeys 

As discussed in previous updates, the impact on migration journeys differs between regions, with increased difficulty moving around inside countries or crossing borders being most frequently cited. Over time, however, we can see an increasing trend of respondents reporting that the crisis had no impact on their journey, see Figure 10, with a proportion of 34% of all participants reporting this in the last period, compared to 20% in the first period. That said, note that this trend is only minor in West Africa (from 7% to 8%).

\

**_Figure 10:  Percentage of respondents who stated that coronavirus had no impact on their journey, over time_**

```{r 10, echo=FALSE, fig.align='left', fig.height=8, fig.width=8, warning=FALSE}
c19  <- data %>% ungroup() %>% #need to ungroup first, which is odd
  select(208, 205, 210, 190:200) %>%#this includes nones
  pivot_longer(cols = 4:14, names_to="Options", values_to="Answer") %>% 
  filter(!is.na(Answer)) %>% 
  group_by(Period, Region, N_period_region, Answer) %>% 
  tally() %>% 
  mutate(Percent=round(n/N_period_region*100, digits = 1)) %>% 
  filter(Answer=="None") %>% 
  ungroup() %>% 
  add_row(Period=1, Region="Overall", N_period_region=2048, Answer="None", n=413, Percent=20.2) %>% 
  add_row(Period=2, Region="Overall", N_period_region=1184, Answer="None", n=278, Percent=23.5) %>%
  add_row(Period=3, Region="Overall", N_period_region=892, Answer="None", n=307, Percent=34.4)

c19$Region <- factor(c19$Region,levels = c("Asia", "East Africa", "Latin America", "North Africa", "West Africa", "Overall"))

fig17  <- c19 %>% 
  ggplot(aes(x=as.factor(Period), y=Percent, color=Region)) +
  geom_line(aes(group=Region, linetype=Region), size=2)+
  scale_y_continuous(limits = c(0, 100), breaks = seq(0, 100, by = 10))+
  scale_linetype_manual(values=c("solid", "solid", "solid", "solid", "solid", "dashed"))+
  scale_color_manual(values=c("#CC79A7", "#009E73", "#F0E442", "#0072B2", "#D55E00", "#000000"))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))
fig17


```
\  
\

This trend is reflected in respondents increasingly reporting that they have not changed their plans as a result of the coronavirus outbreak (from 50% to 57%), while the proportion of respondents who stopped for a while is decreasing (from 24% to 18%), see Figure 11. 
\

**_Figure 11: Impacts of coronavirus crisis on plans, over time_**

```{r 11, echo=FALSE, fig.align='left', fig.height=8, fig.width=9, warning=FALSE}

fig18 <- data %>% 
  group_by(Period, c29) %>% 
  tally() %>% 
  mutate(Percent=round(n/sum(n)*100)) %>%
  filter(!is.na(c29)) %>%
  mutate(c29 = recode(c29, "Yes I have changed my intended destination"="Yes, changed destination", "Yes I have changed my planned routes but my intended destination remains the same"="Yes, changed routes", "Yes I have decided to return home"="Yes, decided to return home", "Yes I have stopped here for a time because I am stuck"="Yes, stopped for a while", "Yes other (specify)"=" Yes, other")) %>% 
  ggplot(aes(x=as.factor(Period), y=Percent, color=c29)) +
  geom_line(aes(group=c29), size=2)+
  scale_y_continuous(breaks = seq(0, 100, by = 10))+
  xlab("Time Period")+
  theme(legend.text = element_text(size = 12))+
  theme(legend.title = element_text(size = 12))+
  theme(axis.text.y = element_text(size=12))+
  theme(axis.text.x = element_text(size=12, face="bold", colour = "black"))+
  theme(axis.title.y = element_text(size=13, vjust= 3))+
  theme(axis.title.x = element_text(size=13, vjust = -2))+
  theme(axis.title = element_text(size = 12))+
  scale_color_brewer(palette = "Paired")+
  theme(panel.border = element_rect(colour = "black", fill=NA, size=0.1))
fig18

```



### Acknowledgement 

The owner and source of the data used for this study is the [Mixed Migration Centre](http://www.mixedmigration.org/), and this study has been published first [here](http://www.mixedmigration.org/resource-type/covid-19/) on 30 June 2020. Data analyst: Jean-Luc Jucker, PhD  
\  
\  
\  
\