HNHacker News
TopNewBestAskShowJobs

houshuang

515 karma · joined April 28, 2011

Post-doc in collaborative learning at EPFL, Switzerland. https://github.com/chili-epfl/FROG https://froglearning.wordpress.com

Engineer / Learning Architect at Minerva Project. https://minervaproject.com

[ my public key: https://keybase.io/houshuang; my proof: https://keybase.io/houshuang/sigs/Y_rXPmmrtVbU1K9cHyNJ7NLuZUHHDd04rA-i7IjA7fs ]

submissionscomments
houshuang··on R, the master troll of statistical languages (2012)
I've just been learning R for my grad studies (and general data exploration), and was really torn between R and Python, especially with IPython Notebook, pandas etc. I came from a Ruby background so both would be newish to me. What made me start with R (although I'm still following IPython dev, and will probably end up switching eventually) is the massive amount of material available to learn stats etc with R.

I've seriously got something like a 100 textbooks on R on my system, intro to stats with R, machine learning with R, questionnaire analysis with R, bayesian stats with R... There are Coursera classes, the amazing r-bloggers.com, etc. For someone who is simultaneously learning statistics and a tool, this is invaluable.

For Python, there's still very little. There's Wes' book, which is mostly about pandas and a lot of finance/time series stuff... I haven't seen a single "intro to stats in social sciences with Python"... There must be one book out there that is open source, which could just be rewritten with Python examples? (The most useful thing I've seen is people reworking examples from Machine Learning for Hackers, or one of the Coursera R courses, in Python).

houshuang··on Ask HN: What are some best resources for learning R?
I've found walk-throughs generated using knitr to be incredibly useful - they show you the commands, and the graphs/tables they generate, they often walk through in small steps so you can follow along, and you also get exposed to a bit of the thinking behind - how to approach analysing different data sets.

My friend just posted this analysis of school board data: http://rpubs.com/dirkchen/knaer-tdsb. RPubs is awesome, you can also browse the front page http://rpubs.com, since all documents are public. Lot's to learn here.

Two things that would make this easier: One is some more social features on the site, like tagging and voting up (show me examples with scatterplots, etc). It would also be great if there was an option to upload the Rmd sheet, as well as the data files when publishing. Of course, some people publish their entire projects to GitHub, but if it was a lot more convenient, perhaps more people would share - one-button import into your own RStudio, for your own experimentation, would be awesome.

houshuang··on Show HN: your personal, self-built productivity tool
It's all open source (although a bitch to install - help or ideas appreciate): http://github.com/houshuang/folders2web
houshuang··on Ask HN: What are some best resources for learning R?
I'll be following this post, since I'm in the same situation (well, in the middle of a graduate program, having done quant classes with SPSS, which I could never grow to like, and now discovering and learning R for my research). There's a lot of great free resources out there - I've been going on a bit of a resource hunt, and gathered lots of links here: http://reganmian.net/wiki/R. Totally unsorted dump at the moment, I'll eventually sort through it, but might still be useful.

Yesterday I spent the whole day watching a Google tech talk by Hadley Wiggins (creator of GGPlot) talking about exploratory data analysis with ggplot and plyr, and replicating every single step in RStudio with RMarkdown (mix of code and text). I played with every single line, took arguments apart, tried out variations, until I understand what each piece did.

Here are my notes: http://rpubs.com/houshuang/ggplot-notes (RStudio makes it ridiculously easy to publish these kinds of documents online). There's a link to the talk, and the data, there too.

One thing I'm looking for is more information about doing survey analysis - I'd love to see a detailed walkthrough (like the file I posted) of someone designing a survey with Likert items, pilot-testing Likert items and generating scales, importing the whole survey with codebook, testing validity of the scales, doing some analysis of findings etc. I've found bits and pieces, but never a complete walk-through.

houshuang··on Show HN: your personal, self-built productivity tool
Here are a bunch of screenshots of my self-built academic productivity system - focused on reading PDFs, taking notes, extracting to a wiki, working on notes etc. http://reganmian.net/wiki/researchr:screenshots
houshuang··on Draft. Version control for writing
Totally agree with you on academia needing this. I know a bunch of people (including myself) who are interested in trying new ways of authoring papers, theses etc - whether it's MarkDown, latex etc. But once you need to collaborate with someone else, or even receive feedback etc (which as a grad student is pretty much always), we just end up back in Word.

I think building something on the backend of Git would be awesome. My prof is pretty computer savvy and open-minded, but I am not going to be able to get him to do a pull request for every change... and neither would I want to. You need to view the changes in-document, easily do comments etc. I'm looking forward to seeing people try to crack this nut!

houshuang··on Introduction to Data Science Online Open Course via Syracuse
There have been several data science/analytics courses etc offered based on R lately, on Coursera, Udacity, etc. Have anyone had good experiences with any of them? I would also love to see a similar course based on iPython/Pandas etc.
houshuang··on [dead]
If people have made this mistake, is there a way of getting it out of the public archive (short of shutting down the repository), or will it forever sit there, part of the indelible history?
houshuang··on Programmer Interrupted
Most of the comments are discussing interruptions in general, and could have been posted on any article about programmer working situation etc, but I think some of the most interesting in this blog post is the attempt to measure cognitive/memory load (both retrospectively, in a research situation, and perhaps eventually live).

Imagine something like the ubiquitous sleep trackers for mobile devices which can track your sleep pattern, and both give you feedback on how you sleep, and also select the best time to wake you up within a certain boundary.

Could we imagine something similar for engineers? We already have very rough tools to provide feedback on productivity, like RescueTime, but a context-aware tool that can model cognitive activity based on edits, switching from one function to another, viewing two files simultaneously, git commits, etc, is far more advanced.

Perhaps if you wanted to ask a programmer a question, you could ask for him to be interrupted sometime during the next half hour, and the system would choose the time where it would be least disruptive?

And I'm also interested in the discussion of ways in which it would be easier to remember/reestablish context. I wonder if some of this is transferable to other domains, for example deep academic "knowledge work" - working with annotated articles, several windows of notes, trying to synthesize ideas etc.

houshuang··on In memory of Aaron Swartz: a collection of PDFs from PDFtribute
Nice short article: 10 things you can do to really support Open Access: http://phylogenomics.blogspot.de/2013/01/10-things-you-can-d...
houshuang··on In memory of Aaron Swartz: a collection of PDFs from PDFtribute
Let me know if you want to discuss further - I've lot's of ideas, but not able to make most of it happen by myself. For example, here are a bunch of (unfinished) notes about an open alternative to Google Scholar, and open social API to share citations/notes, etc. http://reganmian.net/wiki/ideas_for_scrobblr
houshuang··on In memory of Aaron Swartz: a collection of PDFs from PDFtribute
An important thing to remember, is that many journals already permit self-archiving of publications (ie. uploading a pre-print to a personal server or an institutional repository). In fact, about 70% of large publishers automatically allow some form of self-archiving, and for the others, many have been successful including a copyright addendum with the copyright-transfer document, retaining some rights (http://scholars.sciencecommons.org/). FAQ on self-archiving (http://www.eprints.org/openaccess/self-faq/).

At my university, we keep running workshops, there are student staff in the library willing to help upload articles to the repository if you just e-mail them, etc, but still, most academics won't take the five minutes to do this, even if they have the right.

This doesn't mean that the academic publishing system shouldn't change, it absolutely should. And there's also a lot of value in "liberating" academic publications that would otherwise not be free. But I hope people would become more aware of what is already possible, and legal!

houshuang··on In memory of Aaron Swartz: a collection of PDFs from PDFtribute
For someone who has been working on OA and scholarly publications for a few years, it's a bit tricky to enter these kinds of debates. On the one hand, I want to respect Aaron's legacy, and am very touched by the spontaneous and organized moves to honor him. On the other hand, I am very interested in these issues, and love discussing them, seeing how we can do things better.

For smogzer, there is some interesting research on how knowledge from the research literature can be better represented. For example using concept mapping, see this great paper by Simon Buckingham-Shum (who has many others): http://oro.open.ac.uk/6463/1/kmi-04-28.pdf. Anita de Waard has given many presentations on semantic and executable papers, for example http://www.slideshare.net/anitawaard/executing-the-research-....

houshuang··on In memory of Aaron Swartz: a collection of PDFs from PDFtribute
As long as these PDFs are exposed publicly (and linked to, which a tweet with or without #pdftribute will take care of), they will mostly be indexed by Google Scholar, which does a decent job of extracting metadata using heuristics etc.

Of course, it would be much better if people started embedding machine-readable metadata in PDFs (totally possible, see for example http://code.google.com/p/pdfmeat/), and if there was some agreed-upon format for bibliographic microformats, that could be embedded in websites listing articles.

We also eventually need an open alternative to Google Scholar. GS is great, and I use it every day (and love that you can output BibTex for example), but it has no API (and will never have one because of deals with publishers), actively resists automatic access, is a black-box in terms of how data is gathered, etc. Think of "Open Scholar" to Google Scholar as analogous to OSM vs GMaps. OSM might not look as pretty, or be as consistent in the beginning, but it enables a whole range of applications that GMaps doesn't. (And at least GMaps does have a fairly good API, even if it charges for overuse, GS has nothing).

(These are just some thoughts I've made, as I've been experimenting with an open scholar workflow, trying to share as much of the "byproduct" of the research, including rich notes and summaries, my own bibliography with links to OA pubs where they exist etc: http://reganmian.net/wiki/researchr:start).

Another thing I've found working on my project, where I try to expose OA links to as many pubs as possible, and regularly rescan to see if they are still available (and still OA), is how quickly documents disappear... Hosting on private pages is convenient, but fragile. Ideally, people would upload papers to university repositories, subject repositories like Arxiv.org, etc.

houshuang··on Learn to read a binary font by reading a story
I spent quite a bit of time on the first link, but I also quite like this one. I think to me a lot of the value is just gradually introducing a few characters at a time - not really sure about how much my brain is picking up the shapes of the letters subconsciously. Would be interesting to do an experiment.

I would actually prefer that the hints stayed for all the letters, even without a grey line or whatever. I used it very rarely (and have enough self-discipline not to overuse it) but once in a while when I've forgotten a letter, or are not sure, it's a pain to look it up because the hint has disappeared.

It would be fun to experiment with putting this font on my Kindle btw.

houshuang··on Learn to read a binary font by reading a story
Not sure if this is worth the effort, but brilliant approach to learning. I've often had fun playing with inventing my own alphabets and stuff -- if you could actually learn to read this font at a decent speed, it would be fun to click a bookmarklet anytime you wanted some privacy :)

I love reading, and I read pretty fast - I've often thought that the existing approach in most textbooks where they give you a short (5-10 line text) with tons of new words and grammar, and then when you're done working through that, give you a new one with a ton of new words, is non-ideal. Once I'm done learning all the new words, you should give me five pages with text that only uses the words I've already learnt, to "fix" it - and to give me a feeling of mastery...

houshuang··on Show HN: Native Analytics (work in progress)
Also doesn't track when people go directly to download your files, for example finding PDFs through Google Scholar.
houshuang··on Show HN: Native Analytics (work in progress)
Another advantage of "native" analytics is to capture downloads of all file types. Google Analytics does not know about your PDF downloads (unless there is something I don't know about), and or me that is quite important.

I hope you figure out how to block referrer spam. I use both GA and awstats as provided by my webhost, and the numbers are widely different. The referrer list at awstats is almost useless since I get hundreds of hits from viagra selling sites etc, despite the fact that I don't publish the referrer list publicly anywhere.

← PreviousPage 3 of 3