As you may (or may not) know, in addition to hosting public code challenges on HackerRank, we also frequently host or sponsor what we call hackathons. Public, online events where coders from around the world complete the same group of challenges and the top finishers win prizes and swag. In early December we hosted one such hackathon with our friends over at Quora, an event they slyly referred to as their ‘haQathon’. Clever, right?
The Quora HaQathon wasn’t just a sprint, and it wasn’t just a marathon – it was both! The first phase, on December 6th, was an eight-hour code sprint with nine problems. The second phase, that ran from December 7th to Dec 14th, was a seven-day marathon with the same problems. Each phase had its own unique top-25 leaderboard (to which we’ll refer to as ‘the winners’ from here on out). Lots of cool prizes were on the line, too, including the top prize that gave the first place winner the opportunity to meet Quora CEO Adam D’Angelo. Throw in a few brand new iPhone 6 devices and some Quora swag, and we had a fun little pot of prizes to compete for in addition to global bragging rights.
The interest was phenomenal as our haQathon reached over 8,200 signups and around 5,000 participants came out to compete from all across the globe with a very positive reception overall. You can see the breakdown of the ten countries who had the most competitors here:
Numerically speaking, this looks pretty overwhelming in favor of India and the United States which account for 75% of total competitors. By now you’re probably assuming that India and the United States took home all the top honors, too, right? Well, that would be false. After those 5,000 competitors submitted 18,000 total code challenge solutions across both haQathon phases we ended up with some very intriguing data. As with all things HackerRank, it all came down to one thing and one thing only: each competitor’s skill.
Before we dive into how competitors from each country did, let’s examine some of the foundational elements of this haQathon. Starting with the success rate for each of the problems that were posed, we can see that problems P2 [Schedule] and P9 [Sorted Set] were the easiest, and problems P7 [Labeler] and P8 [Duplicate] were by far the most challenging. Another small surprise, was that problem P1 [Archery] (which was rated ‘easy’) was actually more inline with problem P4 [Related Questions] (rated ‘difficult’) and problem P5 [Ontology] (rated ‘advanced’). You can see all of problems here, and check out the full breakdown of how competitors did on each problem below:
Now let’s talk about some of the winning categories. First, we’ll peek at the submission percentage by country. What this analyzes, quite simply, is what percentage of participants from each country submitted at least one solution to the code challenges (we’ll refer to this group of participants as ‘finishers’). What we see here is that South Korea, Latvia, and Hong Kong had the top three spots for completion percentage from countries who had more than ten finishers. Despite huge turnout numbers, India and the United States did not have high completion percentages.
Next, we can also look at the average total score from the top finishers in each country. We don’t want to dilute the average score with data from competitors who didn’t actually finish so we’ve pruned this down to the total average score for just the top 25% of each country’s finishers. Here, we can see that although total competitors and completion rates were higher for other countries, Belarus, Indonesia, Russian Federation, and China rounded out the top four performers by country despite only having 13, 11, 31, and 44 finishers per country, respectively.
Lastly, let’s look at what everyone really cares about: the total number of winners by country. We had nine different countries finish with two or more winners and another 10 countries had a single winner. This means that 19 different countries had a winner in this haQathon. This graphs shows all nine countries with more than one winner:
Admit it: you weren’t expecting to see Belarus claim top honors as the country with the most winners? But that’s exactly what happened and we find that both exciting and insightful. What this means is that four countries in the graph (Belarus, Hong Kong, Japan, and Czech Republic) represent 34% of the winners despite having less than 1% of total competitors. Conversely, India and the United States, despite having 75% of total competitors, only represent 24% of the winners. Add that to the fact that 19 countries had at least one winner, and that’s a pretty bold statement about how hackathons level the programming playing field for everyone.
And that is what all of these statistics truly represent: how HackerRank and hackathons are flattening the coding world. Anyone, from anywhere, can signup, compete, and win one of our hackathons. All they need is a computer and an internet connection. Now, that may sound simple, but in a field that’s been dominated by ‘where you live’, ‘who you know’ or ‘where you went to school’, this flattening of the programmer world is actually quite revolutionary and unprecedented. And, it represents one of the fundamental missions of HackerRank: give anyone, anywhere the opportunity to show off their coding chops and be rewarded for it. These results are proof that we are propelling this revolutionary mission forward one hackathon at a time.