235 points by webmetricshq 5 months ago flag hide 19 comments
bluesky 5 months ago next
Great job! Real-time web analytics is always a hot topic on HN. 99% accuracy is impressive.
justinmind 5 months ago next
I agree, that's a really interesting approach to real-time analytics with high accuracy.
logiclog 5 months ago next
May I know, how did you ensure the accuracy was consistent, especially in real-time?
designer 5 months ago next
We spent a lot of time optimizing the ETL process for real-time data streaming and accuracy monitoring. We use machine learning techniques to automatically detect and handle outliers in the data.
bitand 5 months ago prev next
Amazing work. I have a few questions, what tech stack did you use to achieve this?
stormbreak 5 months ago next
We used a combination of modern data processing tools and real-time analytics libraries, along with some custom in-house work for this project.
cryptobot 5 months ago prev next
Real-time analytics is essential for modern websites and applications. Keep up the good work.
helloworld 5 months ago prev next
Real-time analytics is a great tool, but making sure it is accurate is crucial.
hashbang 5 months ago next
Indeed, 99% accuracy in real-time can make or break a product. How did you measure accuracy?
beyondtime 5 months ago next
We used a combination of synthetic and real user data for testing and verification. When it comes to measuring accuracy, we compared our results to reliable external measurement services to ensure the data's validity.
ryans 5 months ago prev next
Fifty-one percent accuracy was good enough for 'deep learning' AI a few years ago ;) 99% is impressive for real-time analytics!
shanem 5 months ago next
I am glad to hear that a.i algorithms with 51% accuracy made headlines a few years back. This project operates at a whole new level of accuracy and quality.
castellans 5 months ago prev next
Do you use any other technologies like Elasticsearch for querying?
eliben 5 months ago next
No, Elasticsearch is an excellent tool, and we would have considered it. However, for this project, we opted to use a combination of Apache Kafka and ClickHouse for real-time query processing.
ngraph 5 months ago prev next
In the demonstration, you showed analytics data for a webpage. Did you consider using this on other platforms like mobile devices?
vbentley 5 months ago next
Great question! We have considered the idea. It's essential to deliver the same level of accuracy for mobile devices, but we still need to explore the best ways to do this efficiently.
cloudz 5 months ago prev next
There are a lot of real-time analytics tools out there. If considering usage and projected ROI, what do you recommend as a use case?
nicotraciedo 5 months ago next
For monitoring critical user journeys, website uptime, and other high-traffic events, real-time analytics is crucial. For instance, one use case would be monitoring transactions on an online store or signups on a high-traffic landing page.
hedera 5 months ago prev next
Another use case to consider, one that delves into the power of accurate real-time analytics, would be using them for ad attribution or pricing optimization, especially in a competitive market.