What is the best way to report high confidence incorrect classifications from Google Video Intelligence API?
For example one of our submitted videos has a logo that is 87% confidence but misclassified as the wrong organization throughout a majority of the video.
Related
Our main challenge in Google Analytics at the time is to measure the success of our magazine articles.
The problem is that views grow over time so in any timeframe we always have the older articles overshadowing the newer ones. Sidenote: The same problem occurs for measuring social media post success.
My idea of a solution is to measure the rate by which views on articles grow. An article that has a higher growth of views is much more successful than an older article with more views, but with a lower growth rate.
Alternatively something like "views within the first week(s) of publishing this individual article" would also be a good metric.
Unfortunately to some extent also the growth rates rely on this publishing period of individual articles if we are interested in an eternal high score of articles. But since we are mainly interested in recent articles, growth rate would still give us the desired result of showing the most successful recent articles.
Has anyone dealt with the same challenges and found any solution to this, in best case with Google Analytics?
These examples may help, of which I have direct experience.
In the data layer we included a date of publication for the article and then used this to determine growth. This was taken from the CRM and was relatively straightforwards for the dev team. This was stored as a custom dimension in Google Analytics.
We had nothing in the data layer but instead a I just used the date on which page views started appearing as a proxy for date of publication. Not entirely reliable, and you may want to filter by views >5, or whatever is appropriate, to avoid any hits from editors or staff before a page is visible in the site navigation.
In both cases I was exporting data either to Google Sheets (using for example the Google Analytics API addon for sheets) or BigQuery, where it was relatively straight forwards to identify the first date and then calculate, for example, views per day. In your case it would be having a function which looks at the date of publication + 7 days. You may also be able to achieve this with Google Data Studio or similar dashboarding platform.
I'm trying the handwriting recognition API on MS cognitive service, obviously, the results must be not correct in 100%, so we need the confidence value/property of each word returned by API to mark the results in different color, for example, black for high confidence and red for low confidence. User could realize which word are right and which words might be wrong. I didn't find out any information about confidence on azure service.
I need help~~~
[I work on this OCR API at Microsoft]
Sorry for the delayed response - this is a frequently requested feature and is on our product roadmap. Unfortunately, at this time, I cannot provide a specific timeline on when it would be enabled.
I run a monthly report which tracks session views by region, most popular knowledge articles, deflection rates, most popular product pages, software download stats, etc.
We have a new ELT member who is keen to get into the numbers around our contact centre. As I only look after the support site I need only concern myself with putting together a report which outlines what I feel will be useful information around web traffic. I want the report to be brief, and to highlight 4-5 key metrics.
Please can I have some suggestions for data you think would be useful given the target audience?
So far I am considering:
Deflection rates
Bounce rates.
Time on page
Most popular software downloads.
Global session views year to date.
Any help would be really appreciated. Thanks!
I think those metrics are great. Ideally, the value in the data comes from slicing your metrics with a dimension, ie pivoting. For example, bounce rate as an average means little whereas bounce rate by Content Group or Device Category would be more interesting.
Speaking of Device Category, consider completely isolating the metrics for Mobile vs Desktop+Tablet. Those experiences are so drastically different you'd be doing a disservice to average those metrics together.
Lastly, I'd say this new ETL member should get their own access to GA and learn how to pull the data need. GA now offers machine learning insights that quickly surface relevant drivers in metrics; a static approach to KPI reporting is becoming increasingly obsolete.
I am trying to automate the weekly report. Currently, I am using Google Analytics website to get the data for my report. Sampling level is higher precision.
I tried to get the same data by Google analytics API set samplingLevel as HIGHER_PRECISION. However, I am still getting the sampled data.
For FASTER, Precision Level is roughly 25% whereas for DEFAULT and HIGHER_PRECISION sampling level is roughly 50%.
On Google Analytics website, it says 'This report is based on 100% of sessions'. Can I get the same level of accuracy with Google API? I am using Google Apps script.Response for HIGHER_PRECISION is not matching.
Sumit, the API and the Google Analytics UI are certainly different and similarly the sampling's effect on things is a different beast which must be handled properly to get anything useful out of it.
As was mentioned in the comment, you can achieve high precision unsampled reports by (typically) shortening your date range that you're querying for and then "walking" the data.
To walk the data, you are essentially just gradually incrementing that small date range as you move through the desired data.
The "unsampled reports API" is... well, not the best. Considering that's what they are avoiding giving the end user in the first place, the offering available is not a very good long term or large project friendly solution. I would recommend small date ranges and then doing a data walk.
Happy Coding
There are several solutions to avoid sampling issue in Google Analytics by automating the process of data export for short date ranges.
I prefer this tool, it's pretty simple to use: MadStats.io
The goal flow report on my google analytics account shows some strange sampling behavior. While I can usually select up to a month of data before sampling starts it seems to be different for the goal flow report.
As soon as I select more than one day of data the used data set is getting smaller very fast. At three days the report ist based on only 50% of the sessions, which, according to analytics, comes to only 35 sessions.
Has anyone experienced a similar behavior of sampling although only very small data-sets are used?
Sampling is induced when your request is calculation-intensive; there's no 'garunteed point at which it trips.
Goal flow complexity will increase exponentially as you add goals, so even a low number of goals might make this report demand a lot of processing.
Meanwhile you'll find that moast of the standard reports can cover large periods of time without sampling; they are preaggreated, so it's very cheap to load them.
If you want to know more about sampling, see here:
https://stackoverflow.com/a/37386181/5815149