vincentinverso

Thanks so much for the feedback everyone. It was fun and engaging hearing your thoughts. Just wanted to let future readers know that I likely won't come back and read this thread any time soon, so that I can actually go and get some work done ;) Of course if you'd like to discuss without me, feel free..

Thanks for the interesting thoughts ian0. I honestly think Google has a good shot at all the things you describe with their AutoML tables, _if_ they spend a lot of time on the UX. The problem that I'm guessing most services run into though is how to tune (up or down) the informational/educational aspects of the app depending on the prior knowledge of the user. IMO us hackers don't spend enough time on UX, maybe because of time constraints, or maybe because it's not something that's widely taught at university.

Seems like it! Very nice product.

Maybe initially it involves humans if you are just validating the idea. Nice tool!

Followed you on twitter, though I have to admit I'm not much of a Twitter user! Feel free to email me at the address in my profile if that's easy.

> Then you need to set up systems for extracting any 'insight' from what they have and continuously managing and processing data from multiple sources, and so on and so on.

Agreed, this is quite tricky. The solution I described lends itself much better to one-off analyses, thus limiting its usefulness.

> No one is really sitting with a single csv file open in front of them, thinking "Hmm if only someone would tell me what I can do with this".

Yea, most people don't do this. I think the target user would likely be someone who has an ok idea of what they might want to do with their data, but needs some guidance as to what the latest and greatest techniques can actually do, given what they have. There may be a gap between what they think they can do, and what is actually possible. Or they might discover new things that they didn't think of. Perhaps this is more of a personal frustration that I've had in the past.

The "single csv" issue is a concern I have as well. Many data-science-type problems of course cannot be reduced to a single csv.

Thanks for the feedback!

You could think about it via that lens, however this would still be more "choose your own adventure" than hiring a consultant. On the spectrum of problem solving tools, let's say that at one end you have do-it-yourself data analysis tools (not a lot of hand holding). At the other end you have consulting, where your hand is held pretty tightly. This idea would be somewhere in between, but further toward the do-it-yourself end of things, injecting some hand holding and education into the automation in order to get rid of some of the more painful aspects of the learning curve.

Definitely considering!