Closed jeff1evesque closed 6 years ago
We attempted to create a boxplot:
## boxplot: device vs page views
ggplot(meltdf, aes(x=Access, y=as.numeric(value), fill=Access)) +
geom_boxplot()
However, our visualization is only showing points:
We've verified that all meltdf$value
is numeric
:
> all.is.numeric(meltdf$value)
[1] TRUE
The following are our descriptive statistics:
> summary(meltdf[meltdf$Access == 'all-access',])
Access variable value
Length:1071126 2015.07: 59507 Min. :0.000e+00
Class :character 2015.08: 59507 1st Qu.:4.000e+02
Mode :character 2015.09: 59507 Median :2.276e+03
2015.10: 59507 Mean :4.714e+04
2015.11: 59507 3rd Qu.:2.195e+04
2015.12: 59507 Max. :1.294e+09
(Other):714084
> summary(meltdf[meltdf$Access == 'mobile-web',])
Access variable value
Length:532944 2015.07: 29608 Min. : 0
Class :character 2015.08: 29608 1st Qu.: 3171
Mode :character 2015.09: 29608 Median : 10631
2015.10: 29608 Mean : 41651
2015.11: 29608 3rd Qu.: 28329
2015.12: 29608 Max. :231402053
(Other):355296
> summary(meltdf[meltdf$Access == 'desktop',])
Access variable value
Length:506916 2015.07: 28162 Min. :1.500e+01
Class :character 2015.08: 28162 1st Qu.:3.565e+03
Mode :character 2015.09: 28162 Median :1.124e+04
2015.10: 28162 Mean :5.323e+04
2015.11: 28162 3rd Qu.:2.715e+04
2015.12: 28162 Max. :1.144e+09
(Other):337944
Additionally, we verified the Access
column:
> mean(meltdf[meltdf$Access == 'desktop',3])
[1] 53229.77
a9e04c2: we've converted from boxplots to points, since the scale of each representation was too large, causing the interquartile range to be largely insignificant (at least visually). It may be likely to do a transformation on the elements (perhaps logarithmic). However, then the concept of the boxplot is no longer representative of the data points.
We need to adjust various legend titles.
It seems we also forgot to commit some additional changes prior to merging the earlier PR.
We need to adjust our color
to Access
for the visualization/points-total-access.png
case, since it is redundant to color code, based on the y-axis.
We will create visualization(s) between our
Access
column, and the page views, within our basic.R.