For most Domo connectors, you have the choice of entering in relative dates (today-2) or a specific date (1/1/2020). For YouTube's Basic User Activity, it appears you can only enter in numbers of days ago. Does anyone have a solution for entering in static parameters? In my screenshot example, Videos Published After (Days ago): 136 is the equivalent of pulling 1/1/2020 since today is 5/15/2020. But tomorrow 136 will mean 1/2/2020. I want to be able to automate based on hard dates. That way, I can separate 2020 published dates from 2019, 2018, 2017, etc.
I believe the design of the system is so that you can gracefully accumulate history. As I understand it, with youtube (and all / most social media data) there is a seasoning period where metrics will fluctuate as systems update. So that lag window let's you capture any changes to data that occurred in the last n days.
To split out data by 2020, 2019 etc, you could use a dataflow to split up the data OR if you have access to the data views beta (ask your CSM for details), you can split that way too.
Side question. Why do you want to split your data by year?
Thank you for responding. My organization has published over 1300 videos over several years. The YouTube API has a limit of 500 videos that it will pull at any given time. The past few years we have published 200-400 videos per year. So, if I don't include parameters to divide by publish year, then many videos will be skipped over. I think it may return the top 500 videos wioth traffic and ignore the others. Whereas if I separate by publish date and my method is by year, then I can get every single video. What I have been doing is capturing historical data (f = filter year, v = video publish year):
Basic User Activity - 2019f 2019v
Basic User Activity - 2019f 2018v
Basic User Activity - 2019f 2017v
Basic User Activity - 2019f 2016v
Basic User Activity - 2018f 2018v
Basic User Activity - 2018f 2017v
Basic User Activity - 2018f 2016v
This allows me to capture all videos's filter dates. But, I also want to capture current data in batches without having to adjust the publish dates every day. Again, I'm using the publish videos parameters so that all my videos' data is captured and not just the top 500 videos. Does this make sense?
Raise the question to Support@domo.com, there are best practices for building that historical dataset (I believe they'll direct you toward the feature called backfill?)
If I reword your question, it sounds like you're trying to collect metrics for 'all your videos' ever. I am no social media analyst, but I believe the 'limit to 500' invites users to consider a different segmentation than 'date' to conduct trend analysis.
One of my clients pointed out, that unlike facebook posts which tend to have a short shelf-life, videos can be revisited periodically (like Christmas specials or commercials). The data that would be most relevant to this analysis would not be 'tell me about videos published in 2020 vs. 2018, at that point you're interested in 'videos that were published at christmas' ... or any other segment that would be relevant.
Sorry ... i don't have an answer to your specific question, but I believe the system is fundamentally designed to support a different approach to what you're trying to accomplish.
(note. I am not a social media analyst.)
Hi @jaeW_at_Onyx , you are correct. I am trying to collect metrics for all my videos ever published. I am mainly taking this approach because Facebook, Twitter, Google Analytics, etc. provide data this way. Facebook gives daily stats for videos published 5 years ago even if they are only generating 2-5 views a day. However, what you wrote also makes sense. And maybe this is what I will have to settle for if I want to automate current metrics. Thank you!