Hacker News new | past | comments | ask | show | jobs | submit login
Orange: Data Mining Toolbox in Python (biolab.si)
170 points by kawera on July 3, 2016 | hide | past | favorite | 11 comments



Surely things change over the years... right? Orange has been used widely as a bioinformatics tool at Karolinska Institutet and Baylor College of Medicine for example and as a teaching tool for the students. Also: http://bit.ly/29dGD0b, http://bit.ly/29cROFv, http://bit.ly/29slSQ5.


Is it necessary to use bit.ly here?


He clearly needs referral data to mine


Orange has been very handy on occasion, is great for introductory walkthroughs and presentations. I wish more people understood how important the visualization aspect is.


I used this for my masters thesis a few years back. It was fairly intuitive, you still need to be comfortable coding to dig into what's going on under the hood, but the visuals of all the pieces hooked up really helps abstract away some of the complexity and lets the user focus on how the data flows through the algorithm.


I've actually found this tool to be very convenient for explaining the data mining process visually to laymen.

Does anyone have any experience using Orange? How would you rate it compared to SAS EG or other alternatives?


It doesn't.


It doesn't rate against those tools, to be more clear.


Wow, I did some work on this many years ago. It was adding some visualizations using QGraphicsView, and sponsored by Google Summer of Code. It was my first "real" programming project.

Unfortunately I never had a use for Orange. Even while I was working there, nobody could tell me who uses it. There was some cancer research paper from the same team who wrote the software, I think about identifying which genes cause cancer, but nothing else I could find.


I have had great experiences using Rapidminer community edition which is like this but on steroids. I say community edition because they've started charging for the pro versions and probably future versions so community is still free and open source. I think a lot of people miss out by not using these tools to prototype their pipelines. You can get probably an order of magnitude more experimentation done using one of these visual pipeline creators vs. trying to essentially reinvent the experiment infrastructure yourself in your own code.

https://en.m.wikipedia.org/wiki/RapidMiner


Neat. I always wanted to do a node-based application like this but for statistics and tabular data editing..




Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: