Hacker News Reader

Comment by usehackernews | Hacker News Reader

Hi, I’m Dan.

I’m a product manager that is sick of operations & other PMs asking me to write simple SQL queries or create tableau dashboards.

I’ve seen many friends/colleagues that say they want to learn SQL, but never do.

Yet, companies are looking for talent in all departments to learn these skills for decision making.

Rather than having non-technical people try to learn SQL and tableau, I’m exploring a product that translates human text into simple sql queries to enable data analysis for everyone without needing to learn SQL.

“How many [column name] did we have on [Date]”

This isn’t easy, but if done successfully will be extremely valuable.

I’m primarily looking for a data scientist with NLP experience, ideally around semantic search. However, I’m willing to chat with everyone who wants to!

Respond to this comment and/or Email me: wigrizer[at]gmail.com

View on HN

Hi Dan,

If you have a moment check out our demo of this functionality here - https://www.modlr.co/au/demo/ad-hoc-reporting-demo

We've developed a reporting and business modelling platform which has a unique data model which allows for natural language queries of all the dimensions (equivalent to columns) of a dataset.

Here is our website for more information - https://www.modlr.co

Kind Regards, Ben

Hey Ben - Will give this a look. Sounds interesting. Thanks for sending.

Hi Dan, I'm Jesse. I work in the public sector and have seen the pains of trying to get regular folk to learn even basic SOQL for Salesforce, let alone the struggles of folks trying to change the way they think to write a SQL query.

I believe there's a huge market for bridging the power of a structured query language with non-"techy" folks, much like how we see no-code solutions popular among folks who would have (self-admittedly) never been able to figure out how to wire up two RESTful APIs.

My approach to solving this in the past was to create a custom syntax with a transpiler that took a type of English language grammar and converted it into workable SQL. Where I branched from your intention was actually for my book on data science in higher education -- I wanted to have a language where a researcher could say something like this:

"Who is more likely to drop: a student with circumstance_a or a student with circumstance_b?"

That's a rough approximation, but I think you can see what I am talking about.

Are you looking at the text-to-speech space for people to babble at their computers? Or are you looking for some kind of GUI system where people can plug things together? Or are you looking for a grammar (like my approach)? Would love to chat.

Hi Jesse - It sounds as though we are aligned with the existing problem, and the approach. Very interested to talk more. Mind shooting me an email? (My email is listed above in the parent comment)

Reply on news.ycombinator.com