Graphics Group at LunchinatoRs

# Graphics Group at LunchinatoRs
### Eryn Blagg, Katherine Goode, Haley Jeppson, Xiaodan Lyu, Kiegan Rice, Miranda Tilton, Joseph Zemmels
### 2019/11/08

---

<style>

.remark-slide-content {
  background-color: #FFFFFF;
  border-top: 80px solid #23373B;
  font-size: 20px;
  font-weight: 300;
  line-height: 1.5;
  padding: 1em 2em 1em 2em
}

.title-slide .inverse .remark-slide-content {
  background-color: #FAFAFA;
}

.inverse {
  background-color: #23373B;
  text-shadow: none;
}

.title-slide {
  background-color: #FAFAFA;
  border-top: 80px solid #FAFAFA;
}

</style>

.code-bg-gray .remark-code, .code-bg-gray .remark-code * {
 background-color:#E8E8E8!important;
}
</style>

# Graphics Group

- Working group in the statistics department

- Interested in statistical graphics and computational tools

- Meet once a week

- Presentations from graduate students, faculty, alumni, and industry individuals

- [Website](https://isu-graphicsgroup.netlify.com/): post titles and abstracts of weekly talks (sometimes links to the slides)

---

# Figure 1

---

# Original Figure

---

# Facets and More Variability in Color

---

# Facets and More Variability in Color

### Code

```r
# Set the color palette
color_palette = colorRamps::blue2red(n = 8)

# Create a theme for the figure 1 plots
theme_fig1 <- theme_bw() + 
  theme(strip.placement = "outside",
        strip.background = element_rect(color = "white", fill = "white"),
        legend.position = "bottom")

# Plot
fig1_3years_daily %>%      
  ggplot(aes(x = date, y = value, color = elev)) + 
  geom_point(alpha = 0.5, size = 0.5) + 
  facet_grid(variable ~ year, scales = "free_y", switch = "y") +
  labs(x = "Date", 
       y = "Hourly Values by Plot and Elevation", 
       title = "Variable Plots by Topography 2017-2019",
       color = "Relative \nElevation") +
  scale_color_manual(values = color_palette) + 
  theme_fig1 + 
  guides(colour = guide_legend(nrow = 1, override.aes = list(size = 3, alpha = 1)))
```

---

# Summary Values: Daily Averages

---

# Summary Values: Daily Averages

### Code

```r
ggplot() +
  geom_point(data = fig1_3years_daily,
             mapping = aes(x = date, y = value),
             size = 0.25,
             color = "grey75") +
  geom_line(data = fig1_3years_daily_means,
            mapping = aes(x = date, 
                          y = value_mean,
                          group = elev, 
                          color = elev)) + 
  facet_grid(variable ~ year, scales = "free_y", switch = "y") +
  labs(x = "Date", 
       y = "Daily Values by Elevation and Plot", 
       title = "Variable Plots by Topography 2017-2019",
       color = "Relative \nElevation") +
  scale_color_manual(values = color_palette) + 
  scale_fill_manual(values = color_palette) + 
  theme_fig1 +
  guides(colour = guide_legend(nrow = 1))
```

---

# Summary Values: Weekly Average

---

# Summary Values: Weekly Average

### Code

```r
ggplot(data = fig1_3years_weekly,
       mapping = aes(x = week_date, y = weekly_mean, group = elev)) + 
  geom_ribbon(mapping = aes(ymin = value_min, ymax = value_max, fill = elev),
              alpha = 0.15, color = NA) +
  geom_line(mapping = aes(group = elev, color = elev),
            size = 1) +
  facet_grid(variable ~ year, scales = "free_y", switch = "y") +
  labs(x = "Date", 
       y = "Weekly Values by Elevation and Plot", 
       title = "Variable Plots by Topography 2017-2019",
       color = "Relative \nElevation",
       fill = "Relative \nElevation") +
  scale_color_manual(values = color_palette) + 
  scale_fill_manual(values = color_palette) + 
  theme_fig1 +
  guides(colour = guide_legend(nrow = 1))
```

---

# Heat Map over Time

- Time series plots displaying multiple categories can become complicated
- For comparisons across categories, a "heat map" may be more effective

```r
ggplot(data = soilMoisture_data,
       aes(x = date,
           y = plot,
           fill = soilMoisture)) + #"fill" colors interior of tile
  geom_tile() +
  ... #title, axes, and legend specifications
```
]

Example `geom_tile` output (NAs colored gray):

---

# ggplot GIFs with animation package

- Use physical time as a dimension of your plot (make an animation!)
  - Packages: `gganimate`, `animation`
  - `gganimate` is more robust, but `animation` may be simpler to implement

```r
pltList <- list(...) #list of ggplot objects

animation::saveGIF(movie.name = "file_name.gif",
                   expr = 
                     {
                       purrr::walk(pltList,plot) #plot each element of pltList
                     })
```
]

![](../figures/JoePlots/soilMoisture_perDay.gif)

---

# Figure 2

---

# Original Figure

---

# Scatter plot

```r
library(dplyr); 
library(ggplot2)
d2 <- read.table(
  "../data/2_Bina_two-peaks.csv",
  sep = ",", header = TRUE)
d2 %>% 
  ggplot(aes(x = TCG, y = TCW)) + 
  geom_point(
    aes(color = PixelValues),
    alpha = .7) +
* scale_color_gradientn(
*   colors = c("white",
*              "blue",
*              "red"),
*   values = c(0, .2, 1))
```
]

---

# 2-D density plot

```r
d2 %>% ggplot(aes(
  x = TCG, 
  y = TCW, 
  z = PixelValues)) + 
* stat_summary_2d(
*   bins = 50,
*   fun = mean) +
  scale_fill_viridis_c(
    "PixelValues") + 
  theme_bw(base_size = 15) +
  coord_equal()
```
]

# Correlation matrix plot

```r
library(GGally)
my_bin <- 
  function(data, mapping, ...){
    ggplot(data = data, 
           mapping = mapping) +
      geom_bin2d() + 
      scale_fill_viridis_c()
  }
d2 %>% select(
  TCB, TCG, TCW, PixelValues) %>% 
* ggpairs(
*   lower = list(
*     continuous = my_bin
*   ))
```
]

# Figure 3

---

# Original Figure

---

# Without changing Scales
<img src="../figures/Eryn/Figure03plot01.png" width="812" />

---

# Now with porportions
<img src="../figures/Eryn/Figure3plot02.png" width="812" />

---

# Figure 4

---

# Original Figure

---
# With Modifications

```r
ggplot(X4_data) + 
  geom_hline(aes(yintercept = 0), linetype = "dashed", alpha = .7) +
  geom_vline(aes(xintercept = 0), linetype = "dashed", alpha = .7) +
  annotate("text", y = -4.5, x = 4, label = "(-) yield, (+) weed", alpha = .9) +
  annotate("text", y = 4.5, x = -4, label = "(+) yield, (-) weed", alpha = .9) +
  geom_point(aes(y = yieldLRR, x = weedLRR), alpha = .8, color = "gray10") + 
  ylim(c(-6,6)) + 
  xlim(c(-6.5,6.5))  + 
  theme(legend.position = "none") + 
  scale_fill_manual(values=c("ivory2", "tomato", "yellowgreen", "ivory2")) +
  labs(title = "Yield Response vs. Weed Response to Cover Crop", x = "weed response to cover crop", y = "yield response to cover crop") +
  theme_bw()
```
]

```r
ggplot(X4_data) + 
  geom_rect(data = dat, aes(ymin = xmin, ymax = xmax, xmin=ymin, xmax = ymax, fill = cat), alpha = 0.3) + 
  geom_hline(aes(yintercept = 0), linetype = "dashed", alpha = .7) +
  geom_vline(aes(xintercept = 0), linetype = "dashed", alpha = .7) +
  annotate("text", y = -4.5, x = 4, label = "(-) yield, (+) weed", alpha = .9) +
  annotate("text", y = 4.5, x = -4, label = "(+) yield, (-) weed", alpha = .9) +
  geom_point(aes(y = yieldLRR, x = weedLRR), alpha = .8, color = "gray10") + 
  ylim(c(-6,6)) + 
  xlim(c(-6.5,6.5))  + 
  scale_fill_manual(values=c("ivory2", "yellowgreen", "tomato", "ivory2")) +
  labs(title = "Yield Response vs. Weed Response to Cover Crop", x = "weed response to cover crop", y = "yield response to cover crop") +
  theme_bw() + 
  theme(legend.position = "none") 
```
]

---
# Experimental Option

```r
## modify data
X4_data_mod <- X4_data[complete.cases(X4_data),] %>% 
  mutate(yield = ifelse(yieldLRR > 0, "pos_yield", "neg_yield"),
         weed = ifelse(weedLRR > 0, "pos_weed", "neg_weed")) %>% 
  unite(quad, yield:weed) %>% 
  select(-obs_no) %>% 
  group_by(quad) %>% 
  arrange(yieldLRR, .by_group = TRUE) %>% 
  ungroup() %>% 
  mutate(id = 1:109) %>% 
  gather(measurement, value, -c(id, quad))

## code for plot
ggplot(X4_data_mod) + 
  geom_hline(aes(yintercept = 0), linetype = "dashed", alpha = .7) +
  geom_bar(aes(x = id, weight = value, fill = measurement), position = "dodge", width = 1.2) +
  coord_flip() + 
  ylim(c(-6.5,4))  + 
  scale_fill_manual(values=c("#999999", "#E69F00")) +
  labs(x= "observation", y = "LRR", title = "Weed and Yield LRR") + 
  theme_bw() + 
  theme(legend.position = "bottom") 
```
]

# General Recommendations

---

# How to Ask For Help (With Data)

- Context, context, context!  
- Medium of visualization  
- Goal of visualization

---

# Providing Context

- Having data is great! 
- Numbers alone often don't mean anything
- How can you provide necessary context?

---

# Providing Context

- Describe the study or how the numbers were obtained  
- Variables in your dataset: 
    - Categorical or "Word/letter": What do they mean? 
    - Numerical: What do the numbers represent? 
        - Raw measurements?
        - Percentages or proportions? 
             - How were proportions calculated?
        - Change over time or difference between groups?   
        
   
   
---

# Medium of the Visualization

- Who is the audience?  
    - Researchers in your field 
    - Researchers from other fields  
        - This helps determine how much context you need!  
    - Students 
    - Coworkers 
- Where is the visualization going to be used? 
    - Journal article 
        - Formatting requirements
    - Presentation 
        - Talk   
        - Poster  
        - What is your presentation style?

---

# Goal of the Visualization

- Most data can be visualized in many different ways!  
- What is the purpose of the visualization?  
    - Background information  
    - Example of the data structure  
    - Demonstrating data patterns  
- What do you want the audience to get out of the visualization?  
    - Is there a trend, effect, or ordering you want to emphasize?  
    
---

# Goal of the Visualization

[Examples from Nathan Yau's FlowingData](https://flowingdata.com/)

---

# Goal of the Visualization

[Examples from Nathan Yau's FlowingData](https://flowingdata.com/)
<img src = "../figures/kiegan/flowingdata-comparing-two-years.png" width="80%"/>

---

# Annotating Figures

- Labels and captions are really important!  
- Always include units when labeling axes    
- Add a figure caption which provides context

---

# Labelling in `ggplot2`

```r
iris %>% 
  ggplot() + 
  geom_point(aes(x = Sepal.Length, 
                 y = Petal.Length, 
                 color = Species)) + 
  theme_bw() 
```
]

![](slides_files/figure-html/output_iris-1.png)

---

# Labelling in `ggplot2`

![](slides_files/figure-html/output_iris2-1.png)

---

# Labelling in `ggplot2`

```r
iris %>% 
  ggplot() + 
  geom_point(aes(x = Sepal.Length, 
                 y = Petal.Length, 
                 color = Species)) + 
  theme_bw() + 
* labs(x = "Sepal Length (cm)",
*      y = "Petal Length (cm)",
*      title = "Sepal Length vs. Petal Length in Iris Varieties") +
* scale_color_manual(name = "Iris Species",
*                    values = c("mediumorchid4", "plum", "darkblue"),
*                    labels = c("I. Setosa", "I. Versicolor", "I. Virginica"))
```
]

![](slides_files/figure-html/output_iris3-1.png)