2

I'm attempting to create a multi-project timeline that features various stages of the project in order to organize all of our current projects going forward in one easy to visualize place.

I initially found the timeline() package and was able to semi-create what I was hoping for. However, it appears that package is not very customizable, so I am hoping to complete my timeline in ggplot2() instead.

Here is my fake data:

Phase Project StartDate EndDate  
SD Test1 2015-08-01 2015-08-31  
DD Test1 2015-08-31 2015-09-30  
CD Test1 2015-09-30 2015-11-14  
PC Test1 2015-11-14 2015-12-14  
CA Test1 2015-12-14 2016-08-10  
SD Test2 2015-09-01 2015-10-01  
DD Test2 2015-10-01 2015-10-31  
CD Test2 2015-10-31 2015-12-15  
PC Test2 2015-12-15 2016-01-14  
CA Test2 2016-01-14 2017-01-08  
SD Test3 2016-01-01 2016-01-13  
DD Test3 2016-01-13 2016-01-25  
CD Test3 2016-01-25 2016-02-12  
PC Test3 2016-02-12 2016-03-13  
CA Test3 2016-03-13 2017-01-07  
SD Test4 2015-06-01 2015-06-01  
DD Test4 2015-06-01 2015-06-01  
CD Test4 2015-06-01 2015-06-01  
PC Test4 2015-06-01 2015-07-01  
CA Test4 2015-07-01 2015-07-01  
SD Test5 2015-05-01 2015-05-25  
DD Test5 2015-05-25 2015-06-18  
CD Test5 2015-06-18 2015-07-24  
PC Test5 2015-07-24 2015-08-23  
CA Test5 2015-08-23 2015-08-23

As you can see, each project has 5 stages (SD, DD, CD, PC, CA) listed even though some projects (Test4) will not have a couple of stages as they currently have the same date periods listed.

When I ran my the code with timeline(), this is what I got:

    data$StartDate<- as.Date(data$StartDate)
    data$EndDate<- as.Date(data$EndDate)
    timeline(data, text.size=4)

Result

So as you can see, the phases get overlapped in some situations and I can't adjust fonts or center the text in their respective boxes.

Because of this, I am attempting to move to a ggplot, but can't figure out how to make it similar to this timeline.

So far this is my new code:

ggplot(data,aes(x=data$StartDate, y=data$Project)) +
  geom_line()+
  geom_point()+
  geom_text(aes(label=data$Phase), hjust=0,vjust=0)

Result2

As it currently stands, the xlim is too small to show the entire CA phase of a couple of my projects. Also, I'm hoping to be able to customize the text, reorder it by nearest start date to latest, change xlab and ylab and other graphical changes that I don't believe are possible in the timleline() package.

Thanks for any and all help and please let me know if you need any clarification!

Heroka
  • 12,889
  • 1
  • 28
  • 38
medavis6
  • 843
  • 10
  • 32
  • Do you really need the labels in the plot? Because it would be quite hard with the small intervals you have. Is colored blocks and a legend ok? – Heroka Sep 08 '15 at 15:27
  • @Heroka - Yeah, the colored blocks and a legend would be sufficient. – medavis6 Sep 08 '15 at 15:32

1 Answers1

5

I chose to use segment and not ribbon because it's straight lines and segments are much easier to work with with a discrete y-axis.

#make ordered factor so that sequence in legend matches sequence in plot
data$Phase_ordered <- factor(data$Phase,levels=c("SD","DD","CD","PC","CA"))

p1 <- ggplot(data,aes(x=StartDate, y=Project, color=Phase_ordered)) +
  geom_segment(aes(x=StartDate,xend=EndDate,yend=Project),size=15) +
 scale_colour_discrete(guide=guide_legend(override.aes=list(size=7))) #or legend will be too big
p1

enter image description here

Heroka
  • 12,889
  • 1
  • 28
  • 38
  • Perfect! Thank you so much. One last question, when I included all of my actual projects, I get a graph [like this](https://i.gyazo.com/ad88d11673ec96a9f20c49cc260dba5a.png) where the StartDate becomes fuzzy and overlapped. I tried including `scale_x_date()` but it gave me an error "Discrete value supplied to continuous scale". What am I doing wrong? – medavis6 Sep 08 '15 at 16:17
  • Are you sure date is a Date-object in your dataframe? – Heroka Sep 08 '15 at 16:20
  • Looks like it was one of those classic "missing comma" mistakes that are so aggravating. Thanks for checking back in and for the great solution! – medavis6 Sep 08 '15 at 16:25
  • Now that I have set up my plot correctly, `ggplot()` automatically sorts alphabetically on the Y axis by the name of my projects. I've been unable to find out a way to sort by earliest StartDate of the SD phase as the data is sorted in my data frame originally. This would make the graph more aesthetically pleasing and helpful for users. – medavis6 Sep 08 '15 at 21:32
  • 1
    You need to make your y-variable a factor, ordered by StartDate. For instance: `data$Project_ordered <- reorder(data$Project,data$StartDate)`, and put that as y in your plot. – Heroka Sep 09 '15 at 07:25