Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

The first videos? #4

Open
rdstern opened this issue Dec 9, 2016 · 4 comments
Open

The first videos? #4

rdstern opened this issue Dec 9, 2016 · 4 comments
Assignees

Comments

@rdstern
Copy link
Contributor

rdstern commented Dec 9, 2016

After a lot of discussion we have found a useful starting point for the first videos.

  1. The first will be a sort of "promotional video. It will be general and a bit like a video version of the instructions - from section 7 - when we released Version 0.1. We would like to make it available alongside version 0.2.
    a) We will a) open the diamonds data set.
    b) Explain that it looks rather like a spreadsheet and it behaves a bit like a spreadsheet. But it is just a window onto an R data frame. So most happens "behind the scenes".
    c) Could even change a number in the grid, and show that this has issued a command to change it in the R data frame. (Perhaps show this after the right-click, or before?. Need to discuss that.) Then show it goes to 1000 - and hope there is a plus, or something in R-Instat by then to indicate there is more. Explain that we will later show how much more!
    d) Show the right-click in the grid, and mention insert and delete columns perhaps. Show it on the rows and mention filtering. And show it at the bottom and also there show the viewer.
    e) Mention the organise menu. Also mention that these data are well organised and so ready for analysis. Often we have to prepare the data and this is what the organise menu is all about. We have another presentation where we show what happens when the data are not yet ready for analysis.
    f) Notice, at the top that some columns have a "(f)" after the name. R and hence R-Instat distinguish between categorical columns which R calls Factors and number columns. We can see a bit more about our data by looking at these factor columns. Use the levels/labels dialogue.
    g) Results of our data analysis will go here - on the right-hand side. Sometimes we also want the results to be available for further analyses. We show both these features together. Use the summary dialogue. Show results in output window, and also a new dataset.
    h) Use this to show that the spreadsheet can accommodate more than one sheet at a time, i.e. just like a spreadsheet.
    i) There is another feature in R that is also like a spreadsheet. We go to View > Column Metadata. Here we have one line of information that gives details of the properties of each column. This includes what type of column it is. But we also see that it knows that the summary columns are linked to the original data column. This is a bit like linking data in separate sheets in a spreadsheet - rather than copying.
    j) Now let's show some graphs. And here we explain we are using ggplot2 - and also this is just a taste, because another video looks at the system for graphs in more detail. Then I would like to show just a couple of examples, perhaps first a one variable graph of all the data - also useful for data checking, and perhaps we should put it there too.
    k) Then perhaps a boxplot and then facets with a second factor.

  2. A second could be on our graphics - using ggplot. It will be a bit more instructional. But still conceptual. We assume that a video etc isn't needed to learn how to use a dialogue.

  3. Then one on the Organise menu, but really on coping with messy data. With examp-les of different messy sets. And concluding with the point that if you have (say) 20 sets that are messy in the same way, then transfer the code to RStudio. But often different sets seem similar, but are each slightly different. In that case it just takes longer, but the more visual approach in R-Instat will probably not take

  4. And one on the file menu, including importing from SPSS that includes labels, and possibly discussing a bit about ODK, and maybe importing of gridded data, i.e. showing the flexibility. Mentioning rio, and not hiding problems.

  5. I propose one on the Model > One variable, that can also link to the use of distrplus package. And a second on the probability distributions, that can mention the mosaic package. In each case I would like to contact the developers, so this is also with their agreement and possible input. And (in general) I would like the videos to keep mentioning the key packages. This will sometimes be a single package per video.

  6. Problem and data based. This could become a small series. I would like to point users to some of the interesting data sets in the packages and elsewhere that we can access via the Open from Library. I could imagine here having a small series that we start ourselves - say with the R-packages and then agree that part of our search for funds will be to extend this series with our own examples.

  7. Obviously we also want some videos on climatic, and others on corruption. Again we might have "tasters" ourselves, and use the ideas as part of fund-raising.

On methodology, we initially plan to use Camtasia (with Snag-It for screen capture). We can start with a 30 day free trial. If we buy it is about 200 pounds for the two of them bundled together. This will give the videos in a form that (I think) can go into You Tube easily. I assume they can also easily be transferred to Articulate for interactivity.

@rdstern
Copy link
Contributor Author

rdstern commented Dec 10, 2016

Following further discussions in Canada I describe our proposed method of working, on the videos, at least to start with.

  1. Rachel and Myles will do the first video, which is roughly outlined above. This is partly for the video, and also to test the method of production.
  2. They have downloaded the software recommended by Dave Mills, namely Camtasia and Snag-It. This is a free trial version. If these turn out to be suitable, then one copy of each will be bought.
  3. They will prepare the video, i.e. capture the necessary screens. They will also make a separate sound track, once the capture of the video has been made.
  4. They have found that this two-stage process is better than attempting to do the sound track and video capture together.
  5. The problem with this 2-stage process is to then be able to synchronise the two components. This sometimes means edits to the video. However it is far preferable, because it means that repeated attempts can be made to get a good sound track, and particularly with no "ums" and "aaahs".
  6. They will prepare a written script as the basis for the sound track. This also helps with translation into other languages. The 2-stage process makes the addition of multiple languages a relatively easy process. The first draft of the video with ideas for the sound track will initially be prepared by Roger.
  7. They will also prepare a template for the introductory and closing screens. This template will be one that can form the basis for the series of R-Instat videos.
  8. For this first video - which is broad and introductory, and (to some extent) promotional, they will not consider the incorporation of the video into Articulate. Adding articulate has the advantage that it will permit interactivity. It has the disadvantage that the format is then different, and hence it cannot be saved in youtube in the same way. Hence the addition of Articulate will be considered for a subset of the videos at a later date.
  9. They also propose some "infographics" to supplement the existing help and these videos. We envisage that these might be like the "cheat sheets" used by RStudio. Indeed it may even be possible to get their permission (perhaps) to adapt some of their sheets.
  10. The videos will be accessible through the R-Instat help system. Depending on their size, they may, or may not, be downloaded as part of R-Instat. (We view this as similar to the CAST system for videos, where perhaps there might be a separate option to download these materials.
  11. They also add to the resources for R-Instat. There will be other resources, e.g. CAST, ILRI, good practice book, that are, to some degree, associated with the support for R-Instat.
  12. They will join the github account for discussions on issues concerned with the production of the videos. This will be the same section as is used for the help information on R-Instat. We have started the production of a parallel set of issues. We envisage each video will have its own thread as it is produced.
  13. Once the trial system is operational it will be shared with the AMI team. We propose that this sharing - initially with Patrick and Mike - will start in January 2017. This could start initially by Rachel/Myles doing the video capture and sending a proposed sound track, to be recorded at AMI.
  14. AMI could then start, as we are doing here, with the trial version of Camtasia and snagit. They prepare the sound track and merge it with the video.
  15. At this initial stage an early task for AMI is to sort out who will do the recording. It may, but is not necessarily Patrick or Mike. It needs to be someone with a clear voice, that can be understood internationally, and who is not prone to odd pauses, or "ums" etc.
  16. As the AMI team get more experienced, they can take-over more stages in the production process. In particular they could soon be recording the videos as well as the sound track.
  17. In this process, Rachel/Myles will play a sort of "Danny" role. As with the recent work on the code, it will be good to attempts, e.g. for the sound track, to be checked by someone in AMI, before being sent for comment to Rachel/Myles.
  18. Note that even though Rachel is experienced in producing the sound track, it usually involves multiple attempts before she (and Myles) are happy with it. It is a difficult part of the process to sound natural, clear, interesting, etc, when following a sound track.

@rdstern
Copy link
Contributor Author

rdstern commented Dec 12, 2016

One more piece of useful "research" will be on how smooth we can make the process of viewing the videos in you tube, as well as linking easily from the R-Instat help system. It would be good if we could arrange for the videos to follow one another by default, and also have a facility at the end of them to choose which to watch next.

It might also be useful if the sound track is good enough that the videos remain useful without the pictures, i.e. just from the sound.

And Rachel is keen that the videos are not simply a long screen capture. In addition, when we are talking and the screen is not moving it would be good to have some static images, or powerpoiunt-type text. This has the added advantage that these static screens could be visible for differing lengths of time, and that might be useful for the different languages. Of course, we may also want to translate those screens. I wonder if there will be a way of having that capability without having to have totally separate presentations for each language.

Of course, for the ones that show the menus, and dialogues, I hipe we will have those in different languages, so separate versions may be needed.

Another series of videos could be built from some of the old case studies we had with the original Instat. If we are going to redo them, then I would wish to consider that at the same time as we make a video. For me the first two would be "Throw away your statistical tables" and "Start with the primary data".

@myleskg
Copy link

myleskg commented Dec 17, 2016

Here is a possible script for the first video, with some questions mainly for Roger, but open to comments and thoughts from everyone.

Do we know an animator that could be involved in bringing the description of R-Instat to life? (e.g. Somehow showing 5000 packages being crunched down into R-Instat. Myles likes the idea of a meat grinder. ha ha ha.)

R Instat Overview Script for Comment.docx

Just to add that maybe Ben could advise. On reflection I would rather have an animator point us in the right direction and get "us" started, rather than take a role themselves - unless they are keen on sort of "joining". (This is Roger adding this sub-comment!)

@volloholic
Copy link
Member

volloholic commented Dec 18, 2016 via email

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
None yet
Projects
None yet
Development

No branches or pull requests

3 participants