Launched: Data Science Portal


This feels like people who opted to not use the more One click-approach of hassio gets angry that this isn’t an one click solution for them. Almost everything that’s avaliable for hassio is also available for the other install methods, you just have to do it yourself. That’s what you chose.
This IS the great thing about hassio for many people, that addons help people do things they otherwise would have trouble accomplishing themselves.



That’s fine, but that’s not how the original blog post was written. The original blog post was written implied that in order to use this with homeassistant you needed to use a specially written hassio addon. That’s why people were upset. The blog has since been edited to show that you can use it with the other tools, and tells you what tools you need. This is fine. People remain upset because of how the post initially ignored a massive percentage of the user base and that portion of people were basically told to “jog on and work it out for yourselves” which is a HUGE paradigm shift for the community spirit of this project.

I don’t think this was ever in dispute. But there are also a large middle ground of people who are technically articulate enough to want the flexibility of a non-hassio install, but don’t understand the complexities of communicating with homeassistant, and just want a few pointers in the right direction (and not to be forgotten about) :slight_smile:


can one provide some example/screenshot what this project does?
what are the possible use cases?

thank you.


You can explore the data that resembles your HASS instance to gather information about how to build automations and such.


@balloob It’s seams that it is a handful of people screaming their ass of in this case. Way is hard to understand, maybe jealousy :man_shrugging:t3: However, they represent some 0,001% of all users, don’t let them get to you guys. You are doing a awesome job and we are many people that rely appreciate your hard work!

For those how don’t like it, welcome to contribute instead of complaining. HA is not a right, it’s a service!


Even better, you can actually try out the detective library on Binderhub. You can even upload your own .db if you want and do analysis on it. Try the Usage of detective.ipynb notebook. This is basically the ‘minimal effort’ experience you get with the addon. Cheers


Sorry for my ignorance, still don’t see anything useful yet. It has a lot of potential, I will have to learn first how to use the tool.

I just want to know what real life example you would get out of it?
Is it like the bayesian sensor data, but than much more advanced?


My original use case was for optimising a Bayesian sensor - determining inputs and priors to get the right behaviour from the sensor. Other use cases include creating ML models for classification or regression. The options are many


I understand and support @ReneTode concerns. Home Assistant appears to be taking a new direction, and of course that’s worrying to those who took to the original concept, and those who have built their home around it. HA wasn’t advertised to the average user, and that gave maximum flexibility. That uniqueness is without doubt fading, and it’s proabably no coincidence that it started to occur when the financial backing came along. One can only hope that the project continues to believe in the original concept and those who have contributed so much. This new feature probably doesn’t warrant my comments as I admire seeing new features regardless, but I can’t help but support loyal servants to the project such as @ReneTode when I sympathise with their concerns. Let’s not forget those who have helped take the project this far.


I’ve read this thread from the beginning three times I still can’t understand the reason for this controversy. Maybe I’m just dumb (which is probably true), but I really can not see a reason for all this.


What exactly is this “new direction”?


Sorry have to disagree, if anything the “financial backing” has allowed the project to accelerate.
Every project need to evolve and if necessary take new directions to stay relevant, that is good management.

Every now and again, a feature or change is introduced and some people take exception to it,

Ubiquiti Networks investment in baloob - will take HA from being open source was the cry, yet still open source
HA cloud - not in original spirit of project was the call out, yet its helped lots of people and will help all in the long run with allowing the project to take on full time developers released- will other methods not be supported was the criticism. yet they still are.

Even with all the criticism the developers still choose to keep going forward and at a cracking pace. Just because a feature is released with documentation for one platform does not mean the end of HA for the rest of us who choose anther installation method.

So people calm down and move on,


I have some suggestions for the Data Science Portal that I hope would help new users. First, as a non-Data Scientist, it would be great if you could add a definition of “Data Science” on the home page. Next, I’d recommend showing some examples of what can be accomplished to help convince a potential user why it’s worth looking at. Lastly, I think, a couple of walk through examples would be beneficial.


This will come over time. People who make use of it will share their experience so other can learn from it.

To me the data science portal and how it has been released fits well in the style of Home Assistant as a whole. It is released early so that curious people can give it a go and the documentation will grow over time.

All in all, some critics here might have a point but the energy put into all the ranting would be much better spent by contributing to the project as in threads like this.


Appealing to less technical users, because there’s more of them.


All of the points you made are 100% valid, however I think the only criticism that is warranted is that the project has far outpaced the documentation. I’d venture a guess that 50% of the threads on this forum are for issues that could have been answered by the users themselves with better documentation.


No need to apologise, I respect your point of view. I think long term the project will steer away from the original concept, but I hope I’m wrong, and you’re right. I really do.


I think that a failing of any fast paced project
But it does go both ways, the documentation can be updated by anyone.


I really wonder what makes you think so. is “just” a convenient way to run Home Assistant. No feature so far has been added to Home Assistant that requires you to use nor the cloud for that matter. Everything is possible without it.


Hence why this thread went south so quickly. The blog post says: " All of this is powered by the new JupyterLab add-on" and then we clicked the link, and the guide implied that this was the only way to do it.

Turns out it was an ommision in the linked doc, and it has been rectified, but I think the strength of feeling that this core promise had been broken is evident from the emotions in this thread, and are not helped when the ‘big cheese’ enters the fray and says things like “go and work it out yourself” which further reinforces the fear that other methods won’t be supported in future.