Create custom 'ist687viz' package

[jeff1evesque]

We need to create a custom ist687viz package, intended to reduce the exploratory.R code:

##
## Note: boxplot renders very similar to points, since spread is
##       very large, while the interquartile range relatively
##       insignificant to the spread.
##
ggplot(access.m, aes(x=Access, y=value)) +
  geom_point(aes(fill = Access, color = Access), alpha = 0.35) +
  guides(fill=FALSE) +
  labs(x = 'Total: Access type', y = 'Page views', title = 'Access: Page views vs. Type', color = 'Access') +
  theme(plot.title = element_text(hjust = 0.5))

ggsave(
  'visualization/points-total-access.png',
  width = 16,
  height = 9,
  dpi = 100
)

## points: monthly page views by access (density)
ggplot(access.m, aes(x = variable)) +
  geom_point(aes(y = value, color = Access)) +
  labs(x = 'Monthly: Access type', y = 'Page views', title = 'Access: Page views vs. Access type') +
  theme(plot.title = element_text(hjust = 0.5))

ggsave(
  'visualization/points-monthly-access.png',
  width = 16,
  height = 9,
  dpi = 100
)

## convert wide to long
agent.m <- melt(df[-c(1, 3, 4)], id='Agent')

## barchart: monthly page views by agent
ggplot(agent.m, aes(x=variable, y=value, fill=Agent)) +
  geom_bar(stat='identity') +
  labs(x = 'Year.Month', y = 'Page views', title = 'Agent: Page views vs. Year.Month') +
  theme(plot.title = element_text(hjust = 0.5))

ggsave(
  'visualization/barchart-monthly-agent.png',
  width = 16,
  height = 9,
  dpi = 100
)

Note: the above is only a snippet from our exploratory.R.

[jeff1evesque]

This issue is not a priority for the time being. So, we'll close it, and may reopen at a later date.