Closed GoogleCodeExporter closed 9 years ago
Original comment by amylat...@google.com
on 16 Apr 2015 at 10:45
Stephanie, could you please look into this? I investigated and found a couple
of issues:
- For the END state, it looks like we changed the END state from 'END' to None
in the learner view, but forgot to update the stats check in
https://code.google.com/p/oppia/source/browse/core/controllers/reader.py?name=develop#266
so a bunch of StateHit events are being recorded with their state set to
'None'. I think we need to update the learner view method and migrate these to
MaybeLeave events. (Alas, this bug wasn't caught because we don't yet have
integration tests for stuff involving MR jobs.)
- We need to record leave events for the case when a state has an
'EndConversation' interaction. I think we aren't doing that at all at the
moment. Maybe in those cases we should record both a StateHit event and a
MaybeLeave event.
Original comment by s...@seanlip.org
on 16 Apr 2015 at 11:01
Fixed in
https://code.google.com/p/oppia/source/detail?r=64510eceabf1c1eb063a3cdbb8d7e445
4719440b&name=develop
Original comment by s...@google.com
on 24 Apr 2015 at 2:12
Hi, I am experiencing this problem. Statistics reveals the # of completions or
non-completions, but will not give #s on individual States.
What steps will reproduce the problem?
1. Go to "statistics" in an active exploration
2. Click on any state
3. Even though the top statistic says that I have 12 completions and 159
non-completions, every State says that it has been entered 0 times. (see
attached image)
What is the expected output? What do you see instead?
I was expecting to see how many users had clicked on each state.
What operating system are you using?
Yosemite 10.10.3
Please provide any additional information below.
Are Google Analytics available on Oppia explorations? Or, is there some other
method that I need to activate data collection for an exploration?
Original comment by sba...@google.com
on 18 May 2015 at 11:10
Attachments:
Please check these statistics by EOD and let us know if you are seeing the
correct statistics. All changes have been made.
Original comment by amylat...@google.com
on 19 May 2015 at 6:43
Confirming that this appears to be fixed -- the stats aggregator job on the
corresponding instance had stopped running. Restarting it and letting it
complete a cycle fixed it.
sbarry@: thanks for reporting!
Original comment by s...@seanlip.org
on 20 May 2015 at 1:55
This is fixed. Thank you!
Original comment by sba...@google.com
on 20 May 2015 at 5:02
Original issue reported on code.google.com by
amylat...@google.com
on 16 Apr 2015 at 10:44