Thursday, February 28, 2013

How to Interview a Data Scientist, thanks to Daniel Tunkelang

I like this...  ;-)

How to Interview a Data Scientist

The O'Reilly Strata Conference brings together an incredible community of people working on big data. This year, I decided to do something different for my presentation. Rather than talk about science or technology, I addressed the practical problem of interviewing the candidates to build teams of data scientists.
My slides are below, but I'll summarize the take-aways:
1. Keep it real.
  • Avoid whiteboard coding. Filter with FizzBuzz.
  • Use real-world algorithms questions.
  • Ask candidates to design your products.
2. No gotchas.
  • Gotchas reduce the signal-to-noise ratio.
3. Maybe = no.
  • Bad hires suck. Be conservative.
  • Trust your team.
When O'Reilly makes the video available, I'll post it on YouTube.

How to Interview a Data Scientist | LinkedIn

'via Blog this'

Sunday, February 24, 2013

Gedmatch: a DNA geek’s dream site, thanks to The Legal Genealogist

Gedmatch: a DNA geek’s dream site

Great DNA utility website
It occasionally surprises The Legal Genealogist to realize that not everyone is ready, willing and able to test with every DNA company on the planet, or at least in the United States. Just because doing a full round of tests withFamily Tree DNA23andMeAncestryDNA and others will set you back somewhere in the neighborhood of a grand or more is no reason not to let loose your inner DNA geek, is it?
Oh. Yeah. Right.
There is that little problem of the mortgage and the kids’ school clothes and that doctor bill and…
Sigh… Don’t you just hate it when real life interferes with what you want to do?
So to fill in some of the gaps, at least for autosomal DNA tests, let me introduce you to a wonderful, with tools for genetic genealogy research that carry a whopping big price tag of exactly zero. That’s not a typo. The site is free, though donations are gratefully accepted and anybody who uses the site really should kick in — it isn’t cheap to provide the kind of computing power Gedmatch provides.
The brainchild of Curtis Rogers and John Olson (a distant DNA cousin of mine), Gedmatch offers a range of utilities that make it a little easier to extract every bit of potentially useful information out of your autosomal test results. Autosomal DNA testing, remember, is the kind that works across gender lines so you don’t have to find a direct male line from father to son to son (YDNA or Y-DNA1) or a direct female line from mother to daughter to daughter (mitochondrial DNA or mtDNA2). It particularly helps you identify cousins with whom you can share research.3
To use the site to full advantage, you need to download your raw autosomal DNA test results and match lists from your testing company and then upload them to Gedmatch. Both Family Tree DNA and 23andMe make raw data available — AncestryDNA does not — and directions for downloading are on the Gedmatch site. Gedmatch works with the raw results from deCODEme as well. And you can upload a GEDCOM with your family tree to see if you can identify common lines, using technology to help with traditional paper-trail genealogy too.
Although you do have to provide your email address and the exact name you used with your testing company, you don’t have to use your name publicly. You can be John’s Mom, or CountryCuzzin, if you prefer. You don’t have to make any of your information public if you’d rather not, although the utility of the website is limited if you don’t.
Once you’ve uploaded all the files — and there’s more than one to upload — you’re ready to start playing with the utilities.
First, you can compare your results with the results of all other Gedmatch users who’ve made their results public — no matter what company they’ve tested with:
The results can be sorted by the total amount of DNA you have in common with others, measured in units called centimorgans or cM,4 by the longest segment you share, by the number of generations you’re likely distant from your matches and more. Email addresses of your matches may be listed, but can’t be copied-and-pasted to protect them from harvesting by spammers.
You can choose to see information about selected matches in a chromosome browser. Here’s what my results look like against my two uncles and my aunt on Chromosome 1:
And there are several matrix displays available that will display your selected matches not only as they match you but as they match each other, including a very useful option for displaying estimated distance to the most recent common ancestor.
There’s a genetic distance calculator, a relationship calculator, the ability to triangulate on match results to see how you and your match relate to others, a tool for checking to see if your parents are related to each other, and more.
One of my favorite tools is a quick and easy tool labeled “People who match one person, but not the other …and people who match the same 2 people.” I use it to see others that I have in common with a match, and it quickly produces a chart showing information as to how each of us compares to each common match. The chart does have email addresses, deleted here in this example showing matches I share with an uncle.
There’s more than enough here to satisfy your inner DNA geek. You can phase data — if you and one or both of your parents have tested, this utility will help identify what portions of your DNA came from which parent.
And there are six different options for displaying admixture (ethnicity or deep ancestry) data — and each of the six has options galore. Here’s my admixture under just one of those options:
If that’s not enough, you can even get your very own personal chromosome painting:
For anybody interested in learning more about DNA, or even just playing around with results, this is one cool set of utilities. Kudos to Gedmatch, a DNA geek’s dream site!

  1. ISOGG Wiki (, “Y chromosome DNA test,” rev. 23 Jul 2011. 
  2. ISOGG Wiki (, “Mitochondrial DNA,” rev. 30 Jul 2010. 
  3. ISOGG Wiki (, “Autosomal DNA,” rev. 8 Feb 2012. 
  4. ISOGG Wiki (, “Centimorgan,” rev. 24 Jul 2010. 
Print Friendly
This entry was posted in DNA. Bookmark the permalink.

Gedmatch: a DNA geek’s dream site | The Legal Genealogist

'via Blog this'

Saturday, February 23, 2013

MotionX-GPS Track: Boulder winter hike,sunny/nice! Cleats recommended :-)


Connie O'Dell uses MotionX-GPS on the iPhone and is sharing with you the following track:

Name:Boulder winter hike,sunny/nice! Cleats recommended :-)
Date:Feb 23, 2013 11:35 am
(valid until Aug 22, 2013)
View on Map
Distance:4.96 miles
Elapsed Time:1:53:40
Avg Speed:2.6 mph
Max Speed:6.9 mph
Avg Pace:22' 56" per mile
Min Altitude:5,683 ft
Max Altitude:6,443 ft
Start Time:2013-02-23T18:35:02Z
Start Location: 
 Latitude:39º 58' 14" N
 Longitude:105º 15' 57" W
End Location: 
 Latitude:39º 58' 17" N
 Longitude:105º 15' 55" W


MotionX-GPS Commonly Asked Questions

  1. What is MotionX-GPS?
    MotionX-GPS is the essential GPS application for outdoor enthusiasts. It puts an easy-to-use, state-of-the-art handheld GPS on your iPhone.

  2. Can I use MotionX-GPS?
    Sure! MotionX-GPS can be downloaded from the iTunes App Store.

  3. How can I display tracks in Google Earth?
    Follow the directions on the Google Earth web site to download and install the Google Earth program. Save the attached "Boulder winter hikesunnynice Cleats recommended .kmz" file to your computer. Launch Google Earth, select File, Open, and open the saved "Boulder winter hikesunnynice Cleats recommended .kmz" file.

  4. This email was forwarded to me. Where are the attachments?
    Some e-mail programs do not include the original attachments by default when forwarding an e-mail. In this case, the sender must reattach the original files for them to be included.


Please contact MotionX customer support with any comments or questions.

All the best,

The MotionX Team

US and Foreign Patents Granted and Pending. Fullpower® is a registered trademark of Fullpower Technologies, Inc. MotionX™ is a trademark of Fullpower Technologies, Inc. © Copyright 2003 - 2012 Fullpower Technologies, Inc. All rights reserved.

ver: ; ref-id:

Friday, February 22, 2013

SystemC learning (+UVM, System Verilog, ...) at DVCon
Event news: DVCon 2013 - San Jose, CA
Doulos at DVCon 2013:
 Visit the Doulos team at booth 501
 Tutorial: Migrating to UVM
 Tutorial: SystemC productivity
 Other learning opportunities
DVCon logo
Visit the Doulos DVCon booth for a special offer!

Grab an opportunity to catch up with the Doulos team at 
DVCon 2013
, which starts at the DoubleTree hotel, San Joseon February 25, and be among the first to find out about the forthcoming enhancements to the Doulos SystemVerilog training portfolio.

SystemVerilog Out Of The Box
All conference attendees will qualify for a special $200 discount voucher on Doulos SystemVerilog training, so make sure you register with your conference pass at the Doulos booth - 501! 

TUTORIAL: Lessons from the Trenches:
Migrating Legacy Verification Environments to UVM

All new technologies have end user challenges in terms of migration and/or adoption and UVM is no different. Migrating to the Universal Verification Methodology (UVM) from either OVM, VMM, Specman/’e’, or other verification BCL’s is not always easy. 

UVM logo

John AynsleyDoulos CTO John Aynsley will be speaking in this UVM migration and adoption session on Monday February 25th, in which you will hear the real stories from end users who work 'in the trenches' making this conversion magic happen for their teams as they move to UVM.

If you haven’t moved to UVM yet, this is a session you cannot miss.
Full UVM tutorial session details

SystemVerilog/UVM training and support UVM Resources Register for DVCon 2013
TUTORIAL: Increasing Productivity with SystemC in Complex System Design and Verification

After a decade of evolution, IEEE 1666™, aka SystemC™, is widely used for high level system design description and verification. As the system complexity increases, SystemC is becoming an enabler to build platforms for advanced design and verification techniques.

SystemC logo

UVM logoDoulos Senior Member, Technical Staff, David C Black, will be speaking at this tutorial on Monday February 25th, in which some experienced users and tool developers will share their interdisciplinary use of SystemC in building verification environments that provide early hardware access to software developers. 

More details about this System C session on Monday February 25

Other recommended learning opportunities at DVCon 2013

Wednesday February 27:
 The Finer Points of UVM: Tasting Tips for the Connoisseur
 A Tale of Two Languages: SystemVerilog and SystemC

For more information please call the Doulos team on 1 888 GO DOULOS or email
Copyright 2013 Doulos Limited. All rights reserved.

'via Blog this'