Google Dataset Search Engine | Hacker News Reader

Google Dataset Search Engine

datasetsearch.research.google.com

107 pointsidrisschebak31 comments

Not sure how this is a new launch when it's been out of beta since 2020 [0], though good to see it get more visibility as I imagine it's not currently heavily used.

A comparison between Google's dataset search and OpenDataNetwork (which focuses on public data portals)

https://www.opendatanetwork.com/search?q=salary

https://datasetsearch.research.google.com/search?src=0&query...

Google covers more sources, though usually in the case of public sector info (e.g. salaries) I'd prefer just to see the actual sources and not a bunch of Kaggle matches thrown in.

[0] https://searchengineland.com/google-datasets-search-is-out-o...

Some previous discussion:

2021

https://news.ycombinator.com/item?id=27068551

2020

https://news.ycombinator.com/item?id=22130874

2018

https://news.ycombinator.com/item?id=17919297

I think Google has lost all faith in terms of keeping projects around - especially when they involve data locked into a mildly complex system without a complete migration path out.

I would be wary investing time in learning / using any new products from them.

Curious for those that are reading comments here - 1. Are you users of google dataset search? 2. What other dataset searches are you using?

At my company, we recently (last year) did a crawl of over 600 million tables (TabLib) and released it. We've done some indexing, but haven't released a public search for it yet. If people were interested, we could stand up a service pretty quickly that serves a larger number of data sources (with data directly accessable, rather than behind potential paywalls like google dataset search). We can also attach it to our Data AI products (eg. chatgpt integration with it, to do analysis and plotting with it).

Super curious for feedback and signal for interest in a product like this, so please let me know - reach out here or directly via email if you have interest and why!

Seems like it hasnt crawled any sports-related datasets. At least not for baseball. You can get things like ticket sales but not anything like batting averages.

Really wish you could filter by dataset attributes like language, number of rows, number of columns, etc.

Launch is Ok, What's the sunset date?

All jokes aside, this is a very serious question when it comes to Google products [1]. Imagine having spent 5K on a Jamboard...

[1] https://killedbygoogle.com/

I did. No regrets. They supported it as long as our company wanted to use it. By sunset date we had already decided to not renew the license.

And what do you think I have never shipped a single line of Flutter except having toyed pretty decent apps as side projects? The company that can kill Stadia (had AMD GPUs and servers, had its own hardware controller, bought studios, planning to launch its own titles and what not) can surely kill a mobile SDK for sure without giving it a second thought, like at all.

Only things that should be considered constant at this point are the Google's ad business and everything it gets injected into.

Rest is all mortal.

Flutter is open-source though, are you actually worried about Google dropping it and the community being unable to support it?

For context, Killed By Google lists AngularJS, which received updates through 2021, about 5 years after Angular (v2) was released. Polymer is another open-source framework, killed in 2021, and it's still in 'maintenance mode'.

Next round of promotions

Fairly soon, but they have not yet achieved the business process optimizations that will allow them to retire new products on launch day.

It's really annoying because I actually love Google products but try to avoid them whenever possible so I don't become attached to them in anticipation for when it comes time to kill it off. At least with Microsoft, they only kill off products that people genuinely never use or have a better alternative to; also Microsoft almost always provides an official migration path and gives a really long forewarning. Google has a really serious image problem and needs to address it ASAP.

What do you usually love about Google products?

My experience with them has been poor to mixed.

There's no real pattern, it's just that in some cases I would opt for a Google product because it turns out to be cheaper than the competition or in some cases there was no equivalent.

For me the services I got annoyed with the most were Cloud Print/Domains/Latitude/Reader because they were actually decent. In almost all these cases there were no official Google alternatives provided, at least not immediately; and in the case of Cloud Print there really isn't an alternative other than a VPN connection which is incredibly hacky. It bothers me they are discontinued in the way that they were because it just feels disrespectful.

from my observations there are two reasons that Google products continue 1. they print money 2. features can be added

my favorite google product was google talk and suddenly when it had zero growth as it was perfect. I would have loved to have paid for it to keep it available.

Looks like it has been around in some form for a few years now: https://blog.google/products/search/discovering-millions-dat... hopefully that's a sign it will stick around for a few more.

Thank you for bringing this up! I updated the title accordingly

For me, right now.

Not going to use it because your comment is too real.

Doesn't matter, it's not worthwhile to use this tool. As soon as you become productive with it, incorporate it into your workflows and suggest it to your friends, the inevitable will occur. You'll hear the announcement from Google about its termination. At this point in time, it has even become a self-fulfilling prophecy. Since everyone expects it to be dead soon, none will start to use it, causing it to be dead soon.

The joke is repetitive and low effort. When will you have more original material?

Spamming that meme is particularly ludicrous in this case though, given it's a tool that's already been around for six years. That's probably longer than the average product from any company.

Of the 211 services on Google Graveyard, 85 of them were six years or older. So 40%.

I will grant that it's repetitive, but it's also a solid reason to not use any products Google makes if you want to avoid the churn. And it's funny.

It's a solid reason to not use any SaaS. You might be drinking too much anti-Google koolaid if you think their product failure rate is any different than the rest of the field.

At least with a Google SaaS it won't get suddenly ransomwared^W acquired by Oracle or Big Blue or Broadcom. It will die with dignity instead.

> It's a solid reason to not use any SaaS.

That's also true, and I do.

> You might be drinking too much anti-Google koolaid if you think their product failure rate is any different than the rest of the field.

The rate may not be different, but Google has demonstrated organizational willingness to obliterate SaaS products that have an audience but not a sufficiently large one to suit Google's model of revenue generation.

> It will die with dignity instead.

I don't care how my software dies, I care that it died and how much work it will take to replace.

Well, the alternative for a megacorp like Google is to stop attempting to innovate, either top-down (like Apple aborting its entire car platform) or bottom-up by collapsing under the bloat of never making those hard choices (like many other sleepy giants).

It's not a joke. It is a practical question that seems funny because of the circumstances.

How is it a practical question? You're taking on no long-term dependency by using this service. You'd use it to find a data set you need, and then use that data set. If the service were to go away tomorrow, you're still better off having used it than not having used it.

Your comment was totally generic. Nothing about it was in any way specific to this service. The only specific part was your belief that it had just launched, when it was actually six years old. So it seems a bit hard to believe that you were genuinely interested in using this, and the only thing stopping you is having some kind of a guarantee on a minimum lifetime for the service. What you wrote has a much stronger pattern match to seeing what you mistakenly thought was a new product launch, and trying to get in the first "can't wait for this new product to be killed lol" comment.

Have a look at the links to the previous HN discussions that ChrisArchitect posted on the toplevel. You can see exactly the same kind of generic low effort comments implying an imminent deprectation/sunset/shutdown in the threads from 2020 and 2021. And here we are in 2024.

Quality of comment - is your subjective opinion to which you're entitled to but the fact remains that people use Google products and they disappear.

I shifted from Facebook to that Google+ and it got dead. I started using Google Read and then... Switched to Google Podcasts from CastBox and it got retired.

Now Fitbit dashboard is being retired.

And it's not my subjective opinion because I wasn't the only user using those.

Sure, sure. I'm sure you'd have valuable things to post in threads about those services. But you still haven't written a word about this service, it's just repeating the same completely generic complaint that gets posted on HN literally thousands of timer per year.

Comment is pretty generic, I am not claiming it to be unique but I'd say there's precedent of such product retirements and lots of it and that is why it gets repeated a lot. :)

Reply on news.ycombinator.com