Skip to content
Sign in
Artwork for Data Science Tech Brief By HackerNoon
NewsTech News

Data Science Tech Brief By HackerNoon

HackerNoon

Learn the latest data science updates in the tech world.

Play
Open in
  • 24 episodes
  • Avg 10 min
  • English

Episodes 24

NewestOldestLongest
All4Unplayed4In progress0Played0
Counted on this page — what you have heard stays on this device, so it is not something the list can be paged by.
  • June 20 · 10 min

    I Built an Open-Source Firebase Analytics Alternative Because I Hit 1M Events/Day Once Too Many

    This story was originally published on HackerNoon at: https://hackernoon.com/i-built-an-open-source-firebase-analytics-alternative-because-i-hit-1m-eventsday-once-too-many. After hitting Firebase Analytics 1M events/day cap during a mobile game softlaunch, I built an open-source self-hosted analytics pipeline. Here's how. Check more stories related to data-science at: https://hackernoon.com/c/data-science. You can also check exclusive content about #data-engineering, #game-development, #analytics-pipeline, #self-hosted-analytics, #event-streaming, #event-tracking, #product-analytics, #firebase-analytics, and more. This story was written by: @rawbbit. Learn more about this writer by checking @rawbbit's about page, and for more stories, please visit hackernoon.com. A few years ago I was the data engineer on a mobile game soft launch when Firebase Analytics quietly started dropping events past its 1M/day cap. We didn't catch it for days. That experience pushed me to build Rawbbit — an open-source, Apache 2.0, self-hosted analytics pipeline that lands raw events as Parquet in your own object storage. This is the story of why hosted analytics fails at scale, why I chose NATS + Parquet + BigQuery external tables, and what I deliberately left out.

  • June 20 · 11 min

    Your Redshift Cluster Is Probably Idle 85% of the Time — And You're Paying for All of It

    This story was originally published on HackerNoon at: https://hackernoon.com/your-redshift-cluster-is-probably-idle-85percent-of-the-time-and-youre-paying-for-all-of-it. Your Redshift cluster is probably idle most of the day and billing you for all of it. Check more stories related to data-science at: https://hackernoon.com/c/data-science. You can also check exclusive content about #data-analytics, #data-engineering, #data-management, #redshift-data-architecture, #redshift-provisioned, #serverless-rpu, #cloud-cost-optimization, #redshift-data-sharing, and more. This story was written by: @xavariannabarun. Learn more about this writer by checking @xavariannabarun's about page, and for more stories, please visit hackernoon.com. Your Redshift cluster is probably idle most of the day and billing you for all of it. Here's the SQL query, the breakeven formula, and two real production cases that show exactly when Serverless wins, when Provisioned wins, and when neither is the right answer.

  • June 18 · 4 min

    What the Real Operating Data on AI Agents Tells Me as an Investor

    This story was originally published on HackerNoon at: https://hackernoon.com/what-the-real-operating-data-on-ai-agents-tells-me-as-an-investor. Alexander Kopylkov on why AI agents are already running enterprise operations and what the production numbers tell him as an investor. Check more stories related to data-science at: https://hackernoon.com/c/data-science. You can also check exclusive content about #data, #ai, #ai-agents, #investing, #ai-in-business, #ai-customer-service, #ai-adoption, #ai-integration, and more. This story was written by: @alexanderkopylkov. Learn more about this writer by checking @alexanderkopylkov's about page, and for more stories, please visit hackernoon.com. Alexander Kopylkov, venture investor, finds that AI agents are already running core business functions at scale. Klarna automated 67% of its customer service with a single AI agent, saving $40 million. The remaining 33% of complex cases still required human judgment. Only 17% of companies have deployed agents so far, with 60% planning to within the next 12 months.Kopylkov sees the real investment opportunity in the governance layer that makes agents safe to operate on real business accounts, not in the agents themselves.

  • June 17 · 10 min

    Building Data Quality Into the Pipeline Instead of Cleaning Up After It

    This story was originally published on HackerNoon at: https://hackernoon.com/building-data-quality-into-the-pipeline-instead-of-cleaning-up-after-it. Data quality is a pipeline problem, not a form fix. Learn how developers can enforce quality through profiling, matching, and workflow automation at scale. Check more stories related to data-science at: https://hackernoon.com/c/data-science. You can also check exclusive content about #data-quality, #data-engineering, #data-pipeline, #data-management, #data-validation, #data-governance, #data-profiling, #good-company, and more. This story was written by: @melissaindia. Learn more about this writer by checking @melissaindia's about page, and for more stories, please visit hackernoon.com. Bad data costs organisations millions annually and the damage rarely starts at the form level. It starts deep inside production pipelines where incorrect, duplicate, and inconsistent records silently corrupt every decision built on top of them. This article breaks down how developers can take ownership of data quality through five profiling modes, reference table management, standardization and parsing mapplets, deduplication matching, exception workflow automation, and production scheduling, covering the full pipeline from ingestion to deployment. The earlier quality is enforced, the cheaper it is to maintain.

Showing 21–24 of 24 episodes