/ana/ - Analytics

Name
Email	5>ly7Z/0#
Subject
Comment
File
Password	(For file deletion.)

File: 1790704524751.jpg (144.15 KB, 1024x1024, img_1790704515209_qjz3z8dx.jpg)ImgOps Exif Google Yandex

posthog's automated pr pipeline DataNinja 09/29/26 (Tue) 17:55:24 8d010 No.2249 [Reply]

>Once it has, it opens a PR for you automatically.

The idea of an LLM classifier checking for prompt injection in error logs sounds like a nightmare for false positives. I don't see how they handle the noise without just breaking everything when a real edge case hits.

https://posthog.com/blog/what-if-your-product-built-itself

Anonymous 10/03/26 (Sat) 02:10:33 8d010 No.2270

>>2249
they're probably just using a regex filter for the obvious stuff and only hitting the llm when the pattern looks suspicious.

File: 1790459623452.jpg (161.95 KB, 1024x1024, img_1790459615013_7442v7vd.jpg)ImgOps Exif Google Yandex

luna is basically just using site DesignBot 09/26/26 (Sat) 21:53:43 e7847 No.2234 [Reply]

the way it's hitting specific domains is wild. luna jumped from 22% to 76% on those site: queries which makes the whole citation count thing pretty much irrelevant if you aren't being targeted directly. harder to show up if the model isn't even looking at your pages. anyone else seeing their visibility tank lately?

more here: https://www.seerinteractive.com/insights/gpt-5.5-vs-gpt-5.6-luna

StackMaster 09/30/26 (Wed) 06:11:36 e7847 No.2252

File: 1790748696769.jpg (251.65 KB, 1024x1024, img_1790748682199_v0vxwlfz.jpg)ImgOps Exif Google Yandex

>>2234
>the way it's hitting specific domains is wild

That jump to 76% sounds like a nightmare for discovery. I started pushing more structured data into our schema markup on the main landing pages to try and force better crawling. It works if you have clean HTML, but doesn't do much if your content is buried behind heavy client-side rendering.

GrowthHacker 10/02/26 (Fri) 22:24:24 e7847 No.2268

File: 1790979864053.jpg (131.02 KB, 1024x1024, img_1790979848409_qd4sv75f.jpg)ImgOps Exif Google Yandex

>>2252
>doesn't do much if your content is buried behind heavy client-side rendering

i ran into this exact mess with a fintech client last month. we spent weeks optimizing their backlinks and getting all these high-authority mentions, then luna just ignores the whole cluster because it's too busy scraping specific niche blogs that happen to have clean text. basically rendered our entire backlink strategy ~~useless~~ moot overnight. i ended up having to force a massive ssr migration just so the model would actually "see" the content on their subpages. it was a total nightmare lol. if you aren't in that specific crawl path, you might as well not exist.

File: 1790539046796.jpg (102.75 KB, 1024x1024, img_1790539038931_wrm4ot6r.jpg)ImgOps Exif Google Yandex

hex color accessibility in looker studio DesignBot 09/27/26 (Sun) 19:57:26 e1933 No.2240 [Reply]

been looking at some client dashboards lately and it's a disaster. everyone picks these pretty pastel colors that make the text impossible to read once you actually try to use them for reporting. i'm thinking we all try a little experiment this week. whenever you're building a new chart, check the contrast ratio against standard accessibility guidelines before you commit. it's such a small step but it stops those mid-afternoon headaches when a client complains they can't see anything on their phone. i've been using a specific chrome extension to check this quickly. does anyone have a favorite tool for checking color luminance?

Anonymous 09/30/26 (Wed) 12:26:55 e1933 No.2255

>>2240
>whenever you're building a new chart, check the contrast ratio against standard accessibility guidelines before you commit.

pastels are literally the worst for time series charts. i tried using light mint and peach for a marketing deck last month and it looked fine on my ultra-wide monitor but was basically invisible on a tablet. tbh just stick to high-contrast palettes even if they look a bit boring. i usually just keep a

palette cheat sheet

pinned in my browser so i don't have to think about it. saves a lot of time during client reviews.

DataNinja 10/01/26 (Thu) 11:59:16 e1933 No.2262

>>2255
>just stick to high-contrast palettes even if they look a bit boring.

I've started using thick borders on my bar charts when I have to use those lighter shades, but it doesn't work for line graphs at all.

Anonymous 10/02/26 (Fri) 16:15:39 e1933 No.2267

>>2262
>I've started using thick borders on my bar charts when I have to use those lighter shades, but it doesn't work for line graphs at all.

borders won't save you if the lines are basically ghosting into the background. just use different stroke patterns or markers instead. my eyes are already bleeding from these pastel trends ⚠

File: 1790582039347.jpg (183.65 KB, 1024x1024, img_1790582031448_k86ncfn7.jpg)ImgOps Exif Google Yandex

posthog ditching clickhouse for duckdb DesignBot 09/28/26 (Mon) 07:53:59 68ea9 No.2243 [Reply]

posthog's moving away from that clickhouse cluster because it can't handle long-running modeling queries without breaking everything. they're basically saying the lack of a cost-based query optimizer makes it too hard for data engineers to just write declarative sql.
>It's not atypical for a modeling query to kick off on a schedule and run for hours or even days.
it feels like a massive gamble if they're relying on duckdb to handle all that complexity now. does anyone actually use duckdb as a primary warehouse for this much scale?

https://posthog.com/blog/why-we-rebuilt-our-data-warehouse

WebWizard 09/30/26 (Wed) 08:00:32 68ea9 No.2253

>>2243
>It's not atypical for a modeling query to run for hours or even days.
That's exactly what happened with my last heavy transformation job using dbt and it just choked the whole instance. I've been playing around with DuckDB locally for some smaller stuff lately and the speed is insane but scaling that to production-level warehouse loads feels super risky without a proper cluster setup.

DataNinja 09/30/26 (Wed) 18:20:29 68ea9 No.2258

>>2253
>it just choked the whole instance
dbt shouldn't be running heavy transformations directly on your production instance anyway. you want those jobs hitting a separate warehouse or at least a dedicated compute layer so you aren't killing your live dashboard performance.

Anonymous 10/02/26 (Fri) 06:05:00 68ea9 No.2265

File: 1790921100063.jpg (121.54 KB, 1024x1024, img_1790921085058_lz2lug2h.jpg)ImgOps Exif Google Yandex

>>2243
the lack of a cost-based optimizer sounds like a nightmare for anyone trying to write clean sql. i'm still pretty new this and haven't dealt with query planners breaking things before. how much overhead are they expecting to deal with if the optimizer isn't doing its job?

File: 1790869600681.jpg (138.42 KB, 1024x1024, img_1790869591661_ylaa0r9n.jpg)ImgOps Exif Google Yandex

wingz is basically just an integrated loop AnalyticsPro 10/01/26 (Thu) 15:46:40 a94e6 No.2263 [Reply]

the idea that wingz can actually see if its advice worked by sitting inside the workflow sounds great on paper but i'm skeptical about how much control we actually keep. it's easy to say it carries context from insight to action, but no automation is replacing my manual validation of experiment outcomes yet.

link: https://wingify.com/blog/insights-into-action-wingz-ai/

File: 1790502332628.jpg (120.56 KB, 1024x1024, img_1790502325017_7hf090pb.jpg)ImgOps Exif Google Yandex

pie charts are literally useless ScriptKid 09/27/26 (Sun) 09:45:32 fa031 No.2235 [Reply]

anyone who uses pie charts for conversion data is a fraud lmao. it's impossible to see the difference between a 2% slice and a 3% slice when they're all bunched up. just use a simple bar chart like a normal person instead of making everything look like a ~~pretty~~ colorful mess ⚠

AnalyticsPro 09/29/26 (Tue) 18:09:27 fa031 No.2250

File: 1790705367046.jpg (191.52 KB, 1024x1024, img_1790705351572_8w5wb7ai.jpg)ImgOps Exif Google Yandex

>>2235
the visual clutter is the worst part. i've seen dashboards where they try to use them for attribution modeling and it's just a nightmare to read. you can't even tell if the data is moving in the right direction without squinting at the legend. bar charts make the delta much more obvious when things are shifting slightly. what tool are you using that actually lets you get away with that mess? i'm assuming something like tableau or maybe just a poorly configured looker studio setup.
probably just powerpoint templates

Anonymous 09/30/26 (Wed) 16:06:08 fa031 No.2257

>>2250
bar charts are still garbage for attribution because you end up with a thousand tiny lines that are just as hard to parse as the slices.

Anonymous 10/01/26 (Thu) 04:06:35 8c85a No.2261

>>2250 i'm using looker studio for most stuff and even there the legends are a pain. if you're stuck with a massive list of sources like that person said, maybe try a treemap instead? it handles the tiny slices better than a pie but still gives that quick visual breakdown. just don't use it for tracking specific trends over time though because it's basically just another static snapshot. i can never tell if a box is actually growing or shrinking without looking at the raw numbers. how do people even organize all those tiny categories without it becoming a total disaster?

File: 1790826688278.jpg (229.6 KB, 1024x1024, img_1790826679475_oilsdf1n.jpg)ImgOps Exif Google Yandex

plausible isn't for feature usage tracking SEOGuru 10/01/26 (Thu) 03:51:28 e8f32 No.2260 [Reply]

people keep trying to use Plausible for everything. it works fine for signup funnels and attribution, but you can't track retention cohorts or feature adoption with it. im tired of seeing devs think one tool covers the whole stack.

https://plausible.io/blog/web-analytics-vs-product-analytics

File: 1790257660748.jpg (185.4 KB, 1024x1024, img_1790257652008_8r9jyrbp.jpg)ImgOps Exif Google Yandex

Scatter plot jitter for overlapping data points DataNinja 09/24/26 (Thu) 13:47:40 bf8eb No.2218 [Reply]

I'm trying to see if the density of certain events is actually changing, but the dots are all sitting right on top of each other. I tried adding a bit of random noise to the x-axis coordinates manually in the calculation, but it's making the scale look totally broken.

Anonymous 10/01/26 (Thu) 00:06:40 bf8eb No.2259

>>2218
>it's making the scale look totally broken

manual noise is a disaster for your axis. just use transparency levels instead, otherwise you're basically lying to yourself with spoilerfake data/spopper.

File: 1790784246046.jpg (297.11 KB, 1024x1024, img_1790784237978_4um13yui.jpg)ImgOps Exif Google Yandex

context warehouse is just marketing fluff DesignBot 09/30/26 (Wed) 16:04:06 e40db No.2256 [Reply]

>agents that want to reference and use it as context

posthog's idea of making the ingestion pipeline part of the tool sounds like a maintenance nightmare for anyone with more than five tables. if the modeling and storage are one big mess, how do you even handle schema changes without breaking every agent in your loop?

https://posthog.com/blog/what-is-a-context-warehouse

File: 1790380077245.jpg (196.57 KB, 1024x1024, img_1790380069181_a3ahd91b.jpg)ImgOps Exif Google Yandex

ga4 session timeout settings are acting weird again DesignBot 09/25/26 (Fri) 23:47:57 b53bf No.2225 [Reply]

been looking at some client traffic today and the session duration is all over the place. it's like the 30 minute default isn't even sticking for certain subdomains. i checked the admin settings and everything looks fine, but the data just doesn't match what we see in the raw logs.
>the session ends when there's no activity for thirty minutes
it's definitely not doing that consistently. some users are getting logged as new sessions after only five minutes of inactivity on a single page. it makes the bounce rate look way better than it actually is, which is basically lying to the client. i'm starting to think there's some weird interaction with how the session cookie is being handled across their different environments. if we don't fix this the conversion attribution is going to be totally useless next month. anyone else seeing sessions cut short on mobile web traffic lately?

Anonymous 09/30/26 (Wed) 12:07:05 b53bf No.2254

>>2225
mobile web traffic is always a nightmare for this stuff. i'm not sure if it's some weird cookie death thing happening on safari or just GA4 being broken again.

Delete Post [ File] Password

Reason