Histos

• Histograms answer questions related to frequency and data distribution
  1. How similar or diverse are the points in my data set
• Filter Histograms:
  1. Small
  2. Low fidelity
  3. Cue to the user to lead them to narrowing down result set to find specific data points
• Bin size:
  1. Narrow enough to reveal interesting features about the distribution
  2. Wide enough to reduce noise
• AREA is what matters, not height
  

ggplot(hist_data, aes(spanCount)) + 
  geom_histogram(binwidth = .1, col = "black", fill = "blue")

This is test data, but has not been massaged in any way.

Naturalize Histogram by putting all outliers in one bin

hist_data %>%
  mutate(spanCountNew = ifelse(hist_data$spanCount > 10, 10, hist_data$spanCount)) %>%
  ggplot(aes(spanCountNew)) +
  geom_histogram(binwidth = .1, col = "black", fill="blue")

Change Bin Width

ggplot(hist_data, aes(spanCount)) + 
  geom_histogram(binwidth = 1, col = "black", fill = "blue")

Change Bin width and group outliers

hist_data %>%
  mutate(spanCountNew = ifelse(hist_data$spanCount > 10, 10, hist_data$spanCount)) %>%
  ggplot(aes(spanCountNew)) +
  geom_histogram(binwidth = 1, col = "black", fill="blue")