The most interesting thing the Pro version does is to automate the initial steps in exploring your data. It will show you what your data look like and what’s interesting about them, precisely the first steps any data scientist would take upon receiving them. And it’s orders of magnitude cheaper than hiring your own data scientist, at $5/month. The limitations are equally apparent: knowing the right questions to ask, and how to interpret the results, are left to you.

Another incredibly useful feature is the ability to download the numbers behind every graph W|A Pro creates. Raw access to the data enables you to do follow-up analysis, to combine these numbers with others, or to simply make a prettier graph. This turns W|A Pro into a sort of centralized data source — do your search there, and pull the data out to use however you wanted, rather than digging all over the Internet for the data you wanted.

If you’re just getting started with data analysis, I couldn’t recommend a better place to begin. W|A Pro is about democratizing data science. But if the results are unclear, or you’re unsure what conclusions to draw — well, maybe it’s time to ask us. Don’t stop at pretty graphs when you need actionable insights, because they’re all hidden in the data you already have.

No related posts.

I thought, “Could a one-day conference possibly be worth it?” Most of my experiences were with multi-day events like OSCON. At the time, I was on the borderline, mostly because talks I’ve attended have been a real mix in quality, from mind-altering to mind-numbing. But the speakers looked awesome (as they do for Monki Gras), so I decided to go for it. After getting in touch with Steve, I even got invited to give a talk on one of my favorite topics, how to deal with abusive jerks in online communities.

My favorite conferences are the small ones. They’ve got a real community and you get to know everybody instead of stumbling through a crowd of 10,000 strangers. Bigger confs are good for re-connecting with everybody you already know, but small ones are perfect for meeting new people. Somewhere between 50 and 150 people is just about right (Dunbar’s number, anyone? — check when Wikipedia isn’t doing the SOPA blackout). So Monktoberfest seemed like a good fit from a distance, but it remained to see how things worked in actuality.

Finally the day came and I arrived in Portland, Maine. Monktoberfest started with a beer reception the night before at an awesome bar called Novare Res, where I spent the night drinking top-notch beers. You just can’t find this kind of stuff on tap anywhere else, unless you’re thinking of the Lion’s Pride (where we spent the next night). Thursday morning started at a reasonable hour, for a pleasant change, and we had a full day of talks. The talks ranged from “best ever” to merely “outstanding,” so I didn’t even need coffee to keep me awake — just to fend off the caffeine headaches! Lunch featured more excellent beers with a few local selections, as well as lobster rolls since it was Maine, and more talks of an unbelievable caliber followed. At the end of the day, my biggest problem was that my brain was overstimulated.

Thank goodness we had plenty of amazing beer to come. We headed to the Lion’s Pride and had a scrumptious meal along with beer that was so good it was literally unreal. And that’s not even getting into the company; I didn’t talk to one boring or irrelevant person the whole time. Normally, I have to make the rounds just to keep myself entertained. In this case, I had to force myself to move on because I was having so much fun. In fact, the night was so good that one of us bribed the bus driver to stick around longer so we didn’t have to leave! That’s Monktoberfest in a nutshell.

And that’s exactly why you need to come to Monki Gras. Different location (London), different community, but the same RedMonk flavor with a twist of James.

No related posts.

When I was browsing RedMonk Analytics recently, I found some tempting evidence that mobile devices truly are stealing desktop share. But interestingly, this seems to primarily be the case for Windows desktop users, much more so than Mac or Linux. Take a look at this graph of operating systems (OSs) of visitors to redmonk.com, a heavily developer-oriented audience:

The first thing that jumps out at you is that Windows usage drops by 5.2% since mid-2010, down from 61.8% to 56.6%. But where’s that loss going? Those people must be using something else instead, and the question is what. In the same time period, there’s more subtle growth in all of the mobile OSs we measure (iPhone, iPad, Android), for a combined increase of 5.8%. Mobile users now account for nearly 10% of our visitors, tripling from just a year and a half ago (so James, time to set up a mobile theme on your blog!).

To see whether this was a more general trend, I headed over to StatCounter and checked their global stats for U.S.-based web clients. As this is a more general userbase versus the early adopters we see on redmonk.com, you’d expect the trends to lag a bit. But let’s check out the data. First, here’s the graph for mobile vs desktop usage:

What we see is that mobile’s up, no surprise there — it increased from 4.0% to 7.5% since mid-2010. But the real question was whether this disproportionately affects certain desktop OSs over the same time frame:

As you can see, Windows dropped by 4.6% (to 81.2% from 85.8%), while everything else held constant or increased (Macintosh went up by 2.8%), again supporting the trickling flow of Windows desktop use toward mobile devices. Since Microsoft’s still struggling to get traction as a competitive mobile OS, they should be seriously concerned about this threat to its core businesses of Windows and Office.

Of course, they’re already responding in the form of a touch-friendly Windows 8 and a continued push on Windows Phone releases, but things are not looking good at this point. One of Microsoft’s biggest assets is its huge ecosystem of developers running a full Microsoft stack. It should be doing everything it can to push them toward developing for Windows-based mobile devices, because a plethora of apps is a big draw for potential mobile buyers, and the lack thereof will stop them in their tracks.

Disclosure: Microsoft is a client, Apple and Google are not. Linux vendors Red Hat and Canonical are clients.

No related posts.

• Data & analytics. Whether the data’s big, small, or in-between, I love working with it to extract new insights and understanding how others are doing the same. I’m very interested in the ways people cope with and learn from huge amounts of data, whether it’s structured or unstructured. But even on the small scale, there’s a lot you can learn if you apply the right methods. My experience has given me the background in experimental design, data analysis, and statistics to do precisely that.
• IT management, virtualization, cloud, & DevOps. During my 5 years in Oregon, I was closely involved with sysadmins at the Open Source Lab. The OSL hosts servers for major open-source projects including Gentoo Linux (prompting my relationship with them), the Apache Software Foundation, numerous other Linux distributions, and Kernel.org & the Linux Foundation. Interacting with the OSL folks on a daily basis gave me a deep interest in the technologies involved in managing and scaling servers.
• Open-source software. I’ve spent the past 8 years working in OSS projects of all sizes, from 200+ (as mentioned) down to tens (e.g. X.Org) or even just a couple of people (burrow-owl, proteingeometry). Outside of those, I have a broader view of the free- and open-source software world through my participation in the FLOSS Foundations group. I’ve got an insider perspective on how open-source projects operate and look forward to applying that here.
• Community. My positive and negative experiences leading Gentoo Linux and working with the Google Summer of Code have given me a great respect for the power of community and culture. It can make or break companies or open-source projects. For that reason, I plan to spend some time on the cultures of companies, OSS projects, and the communities around them.
• Mobile. It’s pretty clear at this point that there’s an inevitable flow toward open standards and technologies, even within closed app ecosystems. Furthermore, we’re seeing a continuing convergence of more and more functions onto single mobile devices and a migration of functionality to more mobile places. I’m also interested in the re-use of the same technologies in multiple scales and niches (e.g. web technologies like JavaScript and HTML5 used in apps for Windows 8, GNOME 3, and iOS).

That should be enough to give you an idea of my interests, but if you think I might want to hear more about what you’re doing, please do get in touch regardless of whether it’s on the above list.

No related posts.

Who am I, and how did I end up doing this?

James and Steve have already provided ample (and over-the-top!) introductions. So you can understand the background I’m bringing as an analyst, I’ll fill in some of the blanks and the backstory. To sum things up, I’m a scientist and an open-source developer. I started purely in science, until one day in 2001 when I had to learn Linux to do simulations of a neurotransmitter called serotonin (PDF summary), which I was already studying with lasers. I got addicted to Linux, and from there it was just a matter of time — no matter how cool it was to work with lasers.

Turns out that shooting really powerful lasers at serotonin makes it glow purple. (Don't try this at home, kids.)

On the science front, I eventually jumped ship from lasers to X-rays. I worked on understanding the relationships between a protein‘s structure and its function using a number of methods, but the coolest one is called X-ray crystallography. Soon my love of computers became apparent, as I switched gears from doing my own lab work to deriving new insights from large-scale studies of data that already existed (see the pic below), in the form of protein structures solved by others. We then made the tools we created freely available so everyone could benefit. This was the equivalent of so-called “data science” in the world of biochemistry.This graph from my work happens to be about proteins, but it could just as well be any multi-variable dataset:

Using a technique called kernel regressions, we can create nice smooth trends even when the underlying data are pretty noisy. This graph uses color to display the occurrence of different types of protein structure (defined by Φ and Ψ).

My craving to make more of a real-world impact drove me to the Mayo Clinic, where I worked in early-stage drug discovery and continued building my expertise in dealing with large quantities of data. Our favorite technique was called fragment-based screening. We would start computational screening with hundreds of thousands or even millions of compounds, only to narrow it down to ~50 we wanted to test in the lab.

Discovering my passion for technology

At the same time, I became deeply involved in using and programming open-source software (OSS), first just in my free time but later incorporating it into my research. By 2003, I’d settled on Gentoo Linux as my distribution of choice (by way of Red Hat and FreeBSD), and I jumped in with both feet. I started using Gentoo in March, and by June I earned my developer privileges. At the time, I could hardly hack my way out of a paper bag because I lacked any training in computer science. I think I spent most of that summer working on a single package. Fast forward a few years, and you’ll get an idea of how slow of a start that truly was — by 2005, I’d taught myself enough to maintain a few hundred packages, and that was far from the only thing I was doing in Gentoo.

I soon became a leader in Gentoo, first as manager of its desktop project and later as one of the 7 members of its elected council, where I’m serving again after a brief hiatus to focus on science. Gentoo has 200–250 open-source contributors, so getting them all on the same page is no mean feat. A few summers back, I also took over Gentoo’s involvement in the Google Summer of Code (GSoC), an amazing program run by Google that pays college students to work on open-source projects. In GSoC, I oversee ~15 student-mentor pairs; basically, I train mentors, make sure things run smoothly, recruit students to become Gentoo developers, and put fires out.

I write, too!

Outside of science and OSS, my most relevant contribution is probably as a guest author for LWN.net — one of the best developer-targeted sources for open-source news. I was classically trained in journalism in college and spent 4 years working at various newspapers both at the college and professional levels, doing everything from writing to page design to copy editing. Having an opportunity to finally apply this training to my love for open-source development was an incredible stroke of fortune.

Convergence, and joining RedMonk

These three seemingly disparate threads of science, OSS, and journalism have grown increasingly interlinked over the years. At first, I’d wondered how I could possibly choose between my three loves, but somehow things started coming together in a way I’d almost describe as destiny. First it was science and writing, then I started bringing OSS code into my science, and finally I integrated writing into my OSS work with Gentoo and LWN.

Ever since I first discovered Steve 6 or 7 years ago as he was writing about Gentoo, I’ve thought he had the greatest job ever. I’ve closely followed his work over the years, because a lot of it applies quantitative methods to understand trends in technology and how they’re driven by developers, from the bottom up. As a scientist by training, I love quantitation so it was great to see similar methods applied to the IT industry instead of the usual hand-waving. When the opportunity opened up to join RedMonk, I couldn’t resist — between James’ and Steve’s desire to bring on a “data griot,” my admiration for their previous work, and my own diverse background in everything they wanted, it all fit together like the pieces of a jigsaw puzzle.

The hardest part? Making the decision to leave the life sciences behind, even though I love tech and the new directions I’m moving. Between the sheer investment of time and the difficulty of breaking back into academic science once you’ve left, it was nervewracking. But both on the macro- and micro-scales, every other factor made the decision easy — feel free to drop me a line if you’re in the same position. Once I applied, everything went smoothly.

I’ve always felt that the best way to get a job or promotion is to be doing the work already, and James agreed:

But even during the hiring process I was blown away by the fact Donnie just dovetailed with us. If I wrote a post Donnie commented. If Stephen tweeted about some data, Donnie explained how to normalise it. He friended me on every social network I use, and engaged. More than any other candidate Donnie just became a natural part of the team… before we even made the final decision to hire him.

I’m thrilled to be here and am really looking forward to interacting with all of you. In my next post, I’ll discuss my interests as an analyst.

No related posts.