Subscribe
Sign in
FriendlyData — Communicate with databases like a human
FriendlyData is the cloud solution that helps companies make data accessible for non technical people by providing natural language interface to databases.
94
Replies
@mkamenkov thank you, Max! There are a lot of ways how you can use such tool: it can simplify(and replace) complicated interfaces with a lot of radio buttons and checkboxes or increase productivity by adding new functionality to the standard search inputs.
I’d much rather be able to ask questions about the data like a human being, instead of learning SQL or hiring data scientist. FriendlyData makes work with data as simple as searching in Google. Nuff said.
@jeffrey_wyman thank you! You can use it just as it's designed: connect to your database and be able to get any data from it via natural language and not SQL :)
This is a super awesome product, Michael! Any thoughts on adding data visualizations to this?
@javaughn12 thanks for the support! Currently we're focusing on the core API-based technology, that translates english to SQL. We don't provide any UI yet, but we help companies to build advanced dataviz tools in BI products.
@medvednikov thanks Eugene! The idea is natural outgrowth of our experience - we’ve built a lot of sophisticated solutions for big companies. We know that data is the heart of decision making process in any business. At the same time there is no convenient interfaces for non-technical people. This is how we came up with the idea of FriendlyData
@trof thank you, Mikhail! It's a cloud solution, so we provide API and you can build your own shiny UI over it!
Looking forwards to intergration cases on a larger scale! Great product guys! Good luck!
@artuxa for relational database it's size isn't a limit for us.
Thank you, Arthur!
Awesome idea! I wonder how complex your query can get?
For example: Get me customers who have ordered parts shipped from Canada and supplier is not from Canada and orders for last month with transaction more than $500.
And how would you solve DB performance when involved with multiple tables join?
In addition, how would your system know about my table structures? If in my database I have a table called "com_cad" that stores a list of companies name. When I ask "show companies..." how would your system know the companies info is stored in "com_cad" table?
@mingliangma thanks! We support multiple joins but haven't support subquery. It's a good question about DB perfomance. For example, we try to analyze which type of join should we use but it's ofc not a level of good DBA. We do our best to improve query optimizations it in future.
@alexey_zenovich I am really interested in your solution. I filled out the form on your website.
And I'm really curious how would your system know about my table structures?
@mingliangma sorry, didn't see your edit. If the system has some even small data sample it analyzes and generates assumptions about 'human' columns name. Also, it's possible to add like a synonym name by hand and then system analyze queries based on it too.
@dredurr Product is launched in March! You can try our live demo here https://friendlydata.io/demo and request access to our API here https://friendlydata.io/contact
So what you do is basically what Steven Hawking does in terms of explaining physics to ordinary people :)
@aramiggs haha :D truly - our mission is to bridge the gap between technology and human. Thanks for the support!
@csaba_kissi indeed :) Yet we are engineers, we've built FriendlyData to make data friendly for non technical people. Thanks for the support!
You guys have a wonderful product. I think it has a good future in the enterprises. I mean there's a lot of work ahead, but this is something already useful
@eugene_vyborov thanks for the feedback! Agreed, it's just a beginning, but I see a lot of potential on the market and I believe NLP interfaces are going to be mandatory for sophisticated enterprise solutions.
What happens if I misspell a word in the question? Will database still be able to address the question?
@lena_kuznetsova Great questions, Lena! Sure, our system has built-in spellchecker which can handle mistypes and 'fix' them so the database will get a correct question(query).
We're using Postgres. Can we use it? What kind of databases do you support in general?
@arsen_avdalyan yes, you can(and probably you should ;)) use it. Right now we support all popular relational databases, and Postgres is definitely one of them.
How accurate is this and how do you handle things like joins, etc from a simple query?
@allnick Accuracy depends in the query complexity and how big is database structure(bigger database -> harder to provide high accuracy). But it's all matter of time: we can achieve the same high accuracy for big database as for the small one, but it will take more time to teach the system to understand requests properly.
Great questions about joins: right now system require some configuration to be able to handle JOINS, but it need to be done only once.
Hustle X