Google analytics and DSpace

[luked...@gmail.com]

We recently upgrade from DSpace 4 to 6 and we were pleased to get extra reporting such as download events and seeing in real time what is being downloaded.

Our statistics now report 3 months of session day per day and I'm struggling to understand how this can be.

For instance in the past 6 days apparently we have had 100,000 sessions but only 10,000 page views and 150,000 download events. Since the update direct traffic is the most common channel (which seems highly unlikely) and over 99% (or 80,000 sessions) of direct traffic has "(not set)" as the landing page - I assume this is DSpace rather than real people?

The only interpretation I can make is sessions are broken; there are 10,000 page visits and 150,000 files downloaded (probably from Google or another referrer without going to pages).

But I can't really assume something is broken and pick and choose stats to take seriously. Does anyone have any ideas about what is going on?

Thanks.

[Bryson Duda]

I noticed recently that our instance of Dspace (which we also just upgraded to version 6) is behaving similarly with regards to Google Analytics. We're now getting between 90 000 and 110 000 sessions and users per month, and only about 9000 pageviews. Language, Browser, Landing Page, and a number of other dimensions are mostly showing up as "(not set)".

Though the pageviews seem to be comparable to what we were seeing before, the number of sessions are about 25 times higher than we normally see for this instance.

I've tried managing the problem by blacklisting IPs in Google Analytics, but so far it doesn't seem to have helped. I'll be sure to post an update if I manage to find a solution, though.

Bryson

[Neha Singh]

Hi,

I am also using Dspace 6 and I want to know how many files are downloaded per month in Dspace, how can I do this ?

Do file visits= downloaded file count? 

Do I have to write code for this functionality or it is already present in Dspace 6?  

Thanks.

[Bhavesh Patel]

Yes, we are facing the same issue into Google analytics report after upgrading to 6. It seems to be bug or something wrong happen.

Please do the needful on this.

Regards,

Bhavesh

On Tuesday, April 11, 2017 at 5:50:48 AM UTC+5:30, luked...@gmail.com wrote:

[Anne Lawrence]

We just migrated from DSpace 5.8 to DSpace 6.3 on September 11 and have noticed the same huge spike in sessions since then.

Comparing August 16-25, 2019 to September 16-25, 2019:

Sessions: 12,000 vs 197,000

New visitors: 9,000 vs 162,000

Source / Medium    Sessions
google / organic    4,949
(direct) / (none)    2,791
scholar.google.com / referral    1,670
search.vt.edu / referral    297

vs

Source / Medium    Sessions
(direct) / (none)    73,233
vtechworks.lib.vt.edu / referral    62,301
google / organic    43,079
scholar.google.com / referral    6,248

The only major change to our server configuration besides the upgrade is implementing nginx.

Can anyone explain what is happening?

Thanks,

Anne Lawrence

Repository Application Administrator

VTechWorks

https://vtechworks.lib.vt.edu/

[Anne Lawrence]

We are using XMLUI.

[Bhavesh Patel]

Yes, the same issue faced by us also in DSpace 6 version.

Bhavesh

--
All messages to this mailing list should adhere to the DuraSpace Code of Conduct: https://duraspace.org/about/policies/code-of-conduct/
---
You received this message because you are subscribed to the Google Groups "DSpace Technical Support" group.
To unsubscribe from this group and stop receiving emails from it, send an email to dspace-tech...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dspace-tech/89530c9d-b598-4a87-b516-000aaf562e30%40googlegroups.com.

[Bryson Duda]

We're still experiencing the same issue as well with Dspace 6.3 and XMLUI. The workaround that I've been using so far is to create a custom segment in Google Analytics to remove any traffic where the Landing Page is "(not set)". That seems to bring the traffic down to where we would expect it to be prior to the upgrade. I've attached an image of the filter in case it's helpful for anyone. 

I've been digging into this recently, and I think that I might know what's behind it. I believe that Dspace 6 was the first version to include Event tracking in Google Analytics (Neha Singh, I think that this might help answer your question above as well). The idea is that every time a document is downloaded, an event is recorded in GA. If I look under Behavior -> Events -> Top Events, it shows a massive amount of download events (61 000 for Sept. 2019 alone), all of which have Homepage = (not set). I originally wondered if this was a bug with the way that Dspace registers events, but when I checked the Dspace Tomcat logs, I discovered that it does appear to be actual download traffic.

To test this, I picked one document which had 1847 download events listed in Analytics over two days in September, and found an almost identical number of bitstream downloads in the logs for those two days. The downloads came from 610 unique IPs, and there were no other downloads in September after the two day flurry. Geographically, 46% of the download events for that item came from Ashburn, Virginia (which apparently suggests that it likely came from an Amazon data center: https://www.statusbureau.com/analytics-bot-traffic-ashburn/), and the rest were fairly evenly spread throughout the rest of the world. 90% of the download event traffic was direct (not referral, social, or search). Interestingly, the item I chose isn't even a recent upload, it was added to our repository in 2015.

From all of this, my best guess is that the extra traffic is due to direct downloads from bots archiving academic content on the web. The workaround I mentioned above works fine to remove all of the download event traffic from your Google Analytics results, but it would certainly be preferable to only filter out that bot traffic and leave only the download events triggered by real people. I did double-check and the option to exclude hits from bots is already turned on for this view. I suppose for now the best option is to filter out traffic from Ashburn as the article above suggests. If anyone comes across a better solution, I would love to hear it!

Bryson

Bhavesh

To unsubscribe from this group and stop receiving emails from it, send an email to dspac...@googlegroups.com.

[Philip Young]

Thanks very much for this workaround.  We just made the change to GA you describe, and our statistics appear normal again.  Thanks!

Philip

To unsubscribe from this group and stop receiving emails from it, send an email to dspace-tech...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/dspace-tech/689c3caa-9353-47e4-837a-72834530b755%40googlegroups.com.

--

Philip Young

Institutional Repository Manager, VTechWorks

Email: pyo...@vt.edu | Tel: 540-231-8845 | Office: Newman Library 423
Twitter: @openatvt | Blog: Open@VT | Guides: Open Access  ORCID  Perma.cc

Virginia Tech
University Libraries (0434)
560 Drillfield Drive
Blacksburg, VA 24061
http://www.lib.vt.edu