<?xml version="1.0" encoding="UTF-8"?><rss xmlns:dc="http://purl.org/dc/elements/1.1/" xmlns:content="http://purl.org/rss/1.0/modules/content/" xmlns:atom="http://www.w3.org/2005/Atom" version="2.0" xmlns:cc="http://cyber.law.harvard.edu/rss/creativeCommonsRssModule.html">
    <channel>
        <title><![CDATA[Stories by Chris Drouin on Medium]]></title>
        <description><![CDATA[Stories by Chris Drouin on Medium]]></description>
        <link>https://medium.com/@chris.drouin?source=rss-2d753462da94------2</link>
        <image>
            <url>https://cdn-images-1.medium.com/fit/c/150/150/0*xwFj6jMqX6vRWUp1.jpg</url>
            <title>Stories by Chris Drouin on Medium</title>
            <link>https://medium.com/@chris.drouin?source=rss-2d753462da94------2</link>
        </image>
        <generator>Medium</generator>
        <lastBuildDate>Sat, 21 Nov 2026 22:55:06 GMT</lastBuildDate>
        <atom:link href="https://medium.com/@chris.drouin/feed" rel="self" type="application/rss+xml"/>
        <webMaster><![CDATA[yourfriends@medium.com]]></webMaster>
        <atom:link href="http://medium.superfeedr.com/" rel="hub"/>
        <item>
            <title><![CDATA[How Long till the Finish Line?]]></title>
            <link>https://medium.com/runkeeper-everyone-every-run/how-long-till-the-finish-line-494361cc901b?source=rss-2d753462da94------2</link>
            <guid isPermaLink="false">https://medium.com/p/494361cc901b</guid>
            <category><![CDATA[running]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Chris Drouin]]></dc:creator>
            <pubDate>Fri, 25 Aug 2017 13:45:01 GMT</pubDate>
            <atom:updated>2017-08-25T14:07:35.946Z</atom:updated>
            <content:encoded><![CDATA[<h3>How Long Until the Finish Line?</h3><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*B3gjb1lMCb05Tc_OLbMCkA.png" /></figure><p>I’m prepping for a 10K race right now, and I have heard and seen all kinds of misinformation on the internet about what time I should be shooting for / what’s expected. Fortunately, as a data analyst at Runkeeper, I know for a fact that this question can be answered definitively.</p><p>In the charts below, I’ve marked three different times for our male and female race participants:</p><ul><li><strong>5% fastest</strong>: The fastest 5% of our users achieved this finish time or faster.</li><li><strong>Median</strong>: 50% of our users finished faster and 50% of our users finished slower than this time.</li><li><strong>5% slowest</strong>: The slowest 5% of our users finished with this time or higher.</li></ul><p>What you should be shooting for will depend a lot on your own circumstances (age, health, etc.). Don’t be discouraged if your time falls on the right-hand side of these charts: you’re still going faster (and getting healthier) than everyone who isn’t out there running!</p><p>There are other types of races out there, but for now I’ve covered the most typical distances that we see among the races that Runkeeper users participate in: 5K, 5-Mile, 10K, 15K, 10-Mile, Half Marathon, and Marathon. Read on to learn how Runkeeper runners typically finish on these races!</p><p><strong>5K Race Finish Times</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/958/1*cId1fmgkT9sNVtrVvXKcAg.png" /></figure><p>5K races are some of the most popular out there —people at all levels of experience can participate, they get scheduled frequently, and they don’t require a ton of training in advance. Our own Global 5K events have had hundreds of thousands of participants over the years — I highly recommend giving them a shot if you haven’t already!</p><p>The common wisdom that I’ve heard is that 30min is a good time to shoot for for a beginner. That may be a bit aggressive, seeing as the median time for women is 33min and the median time for men is 28min — if you hit 30min you’re already faster than about half of 5K runners!</p><p><strong>5-Mile Race Finish Times</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/958/1*ZqaNKbGGNLnzkxsvqaONIQ.png" /></figure><p>Five milers are less common on the whole — as with their big brother 10-milers, these are more popular in the US than in the rest of the world.</p><p>Finish times for these races have a pretty smooth looking distribution, compared to most of the others we’ve looked at here. I suspect this may be because there isn’t as much “common knowledge” about what times one should aim for. As we’ll see, some of the longer-distance race types have a pretty heavy skew toward particular times…</p><p><strong>10K Race Finish Times</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/958/1*CdCWwHgJ39qBa75w1IV0Ew.png" /></figure><p>10Ks (~6.2mi, for all of us who grew up on the Imperial system) are very popular races worldwide. A lot of our male runners appear to be shooting for a 50 or 55min time, whereas our female runners spike at 60min and 65min.</p><p>In the original discussion that prompted me to investigate this question, I encountered the claim that 45min was a typical 10K finish time. It’s certainly achievable, but as you can see above, it’s likely limited to the top 10% or so of male runners and a small fraction of a percent of all female runners.</p><p><strong>15K Race Finish Times</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/958/1*CIxgRx3cBr6YUrn1HkASkw.png" /></figure><p>Sitting in between the 10K and the half-marathon, 15Ks (9.3mi) are a bit less common as a race distance. This is the first race type where we start to see some runners take more than 2 hours to finish. The bulk of runners finish between 1hr15min and 1hr45min, and the very fastest runners take just over an hour.</p><p><strong>10-Mile Race Finish Times</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/958/1*na2Zp_wtTq8NIohmr4EsBg.png" /></figure><p>10-milers and 15Ks are actually extremely similar races in terms of distance (9.3mi vs. 10mi), so it’s no surprise to see that the two have fairly similar distributions — for most runners, it appears to take somewhere between 5–7 min longer to run the remaining ~680 meters and hit the 10-mile mark. They’re about equal to 15Ks in terms of popularity (which is to say, not terribly common).</p><p><strong>Half Marathon Finish Times</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/958/1*ee_g3SP2XjN_lRB8kA_MPA.png" /></figure><p>Completing a half marathon (13.1mi) is a significant achievement, and many users train themselves to try for specific times. This results in the somewhat lopsided finish time distribution you see above, where the numbers skew hard to the left of the 2-hour mark for men. Having said that, about half of all male half marathon runners are tracking above the 2-hour mark. We see a similar effect for women, but their distribution is smoother overall.</p><p><strong>Marathon Finish Times</strong></p><figure><img alt="" src="https://cdn-images-1.medium.com/max/958/1*G4xfd_2PvRjE4l09VvGvJg.png" /></figure><p>Even more so than with half marathons, we see some interesting cut-off patterns in the data for full marathons. There are large drops in the distribution at the 4hr, 4hr30min, and 5hr marks. I suspect this may be due to a few factors:</p><ul><li>Many marathon training plans, including some older Runkeeper training plans, target specific finish times: 4hr, 4hr30min, etc.</li><li>We’ve heard from Runkeeper users that they sometimes won’t record their finish time if they are not satisfied with it, because they don’t want it to “count” in their official in-app stats. (This breaks my heart as a data analyst! Consider this a plea to record <em>all</em> your runs.)</li><li>Many of the big-name marathons around the world require would-be participants to hit qualifying times. Runners in many age brackets would have to be at or near the fastest 5% to qualify for the Boston Marathon, for instance.</li><li>Marathons can start to push the limits of phone batteries. If you’re headed out for a big race, be sure to charge your phone all the way to 100% first! If your battery <em>does</em> conk out, remember that you can edit and log your activities at <a href="https://runkeeper.com/">runkeeper.com</a> . Runkeeper also works with Garmin and other wearables, so be sure to take advantage of this if you own one.</li></ul><p>I admit to being highly suspicious of some of the times at the extreme left of the graph here — a few of them would be world record breaking if real.</p><p><strong>Bonus analysis: Race popularity</strong></p><p>To give a little context, I thought it’d be helpful to give a little context for the popularity of different race types among Runkeeper users:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1019/1*6a758ibevpGI1cFxr6I2Hg.png" /></figure><p>10Ks and half marathons are the most popular race types by a substantial margin. This is interesting to me, because 5km is the single most common distance that Runkeeper users run. In part, this may be a side effect of the methodology that I’m using to infer race participation: many 5K races are comparatively small / informal events, which may not attract enough Runkeeper users to hit my participation threshold of 25 users.</p><p><strong>Methodology</strong></p><p>How do we know whether a given run is actually a race or not? In some cases, our users tag runs… but for the most part, we had to figure this out with data. In this case, I looked for sets of 25 or more runs that met the following criteria:</p><ul><li>Occurred on the same day</li><li>Occurred at approximately the same time (blocked out in 3-hour chunks, to account for differing start times in some races)</li><li>Traveled a race-like distance (~5K, ~10K, etc.) in a reasonably race-like length of time</li><li>Started around the same location (within a 500th of a degree of lat/lon)</li></ul><p>The charts above were produced using <a href="https://matplotlib.org/">matplotlib</a>, <a href="https://seaborn.pydata.org/">seaborn</a>, and 8 years of race-running data from Runkeeper-using runners around the world. If you don’t already have Runkeeper on your iPhone or Android phone, you can grab it <a href="https://runkeeper.com/">here</a>. Happy running!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=494361cc901b" width="1" height="1" alt=""><hr><p><a href="https://medium.com/runkeeper-everyone-every-run/how-long-till-the-finish-line-494361cc901b">How Long till the Finish Line?</a> was originally published in <a href="https://medium.com/runkeeper-everyone-every-run">ASICS Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[Here’s What “24 hours” of Running in Boston Looks Like]]></title>
            <link>https://medium.com/runkeeper-everyone-every-run/heres-what-24-hours-of-running-in-boston-looks-like-33aefb1f6e89?source=rss-2d753462da94------2</link>
            <guid isPermaLink="false">https://medium.com/p/33aefb1f6e89</guid>
            <category><![CDATA[boston]]></category>
            <category><![CDATA[running]]></category>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[data-visualization]]></category>
            <dc:creator><![CDATA[Chris Drouin]]></dc:creator>
            <pubDate>Wed, 26 Oct 2016 17:20:21 GMT</pubDate>
            <atom:updated>2016-10-26T17:20:21.175Z</atom:updated>
            <content:encoded><![CDATA[<p>My job as Runkeeper’s data analyst has me <a href="https://medium.com/runkeeper-everyone-every-run/a-map-of-the-running-season-7c24242f8b2#.tt4gssspj">deep in our numbers</a> all day long, but lately I’ve been experimenting with geovisualization of Runkeeper runs for fun. I chose to start with runs from Runkeeper employees in and around the Boston area throughout our company history, playing back a few thousand runs as if they occurred in just one day.</p><p>Here it is. (If you want to skip to the action, jump to about the 25-second mark; we like running at Runkeeper but not quite to the point that we’d go running before 5am.)</p><iframe src="https://cdn.embedly.com/widgets/media.html?src=https%3A%2F%2Fwww.youtube.com%2Fembed%2FgIT2hn1dePo%3Ffeature%3Doembed&amp;url=http%3A%2F%2Fwww.youtube.com%2Fwatch%3Fv%3DgIT2hn1dePo&amp;image=https%3A%2F%2Fi.ytimg.com%2Fvi%2FgIT2hn1dePo%2Fhqdefault.jpg&amp;key=d04bfffea46d4aeda930ec88cc64b87c&amp;type=text%2Fhtml&amp;schema=youtube" width="854" height="480" frameborder="0" scrolling="no"><a href="https://medium.com/media/673b21efcbcef455f16aa006470d20d2/href">https://medium.com/media/673b21efcbcef455f16aa006470d20d2/href</a></iframe><p>Each run leaves a faint trail behind it, and as hundreds of trails layer over each other, a map of Boston emerges. This isn’t a perfectly representative map, but it captures a number of the most popular running destinations around Boston, including the Charles River loop, the Chestnut Hill Reservoir, Castle Island, and more. Look closely and you’ll see the Boston Marathon route, with its famed “Right on Hereford, Left on Boylston” turns. All those years of running ultimately wind up looking like this:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*X5c86nrUb2nCqN0IC3Xmpg.png" /><figcaption>The final map of Runkeeper employee runs in and around Boston</figcaption></figure><p>For the curious, here’s the process that went into creating this:</p><p><strong>Identifying the Activities</strong></p><p>I gathered metadata on the runs (“trips,” in internal Runkeeper parlance) that I was interested in looking at with a query to our Redshift data warehouse. At this stage I did some light filtering to exclude activities that were obviously erroneous (traveling significantly faster or slower than normal human running speeds, traveling very short distances, etc.) and bounded my search to trips in the Boston area based on latitude and longitude.</p><p><strong>Pre-processing Path Data</strong></p><p>I then collected and processed the trip path data into a more usable format. Paths get stored as sets of latitude/longitude points by default, which is usually fine — but we need to project those coordinates or the results wind up vertically compressed, as they appear in this example:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*INR0q33ReZeDJQba4kwhhA.png" /><figcaption>Downtown Boston rendered with raw lat/long coordinates</figcaption></figure><p>In this case, I used the Python UTM package to project our lat/long coordinates into the Universal Transverse Mercator system. This is a good quick hack, but doesn’t work everywhere — cities like London that lie on the boundary of two UTM zones are hard to represent with this method.</p><p>Alongside this conversion, I also calculated time offsets — seconds since midnight — for each point in each path. This data goes into creating the animation…</p><p><strong>Rendering the Results</strong></p><p>Having processed the points data, I then needed to render everything. This involves stepping through the seconds in the day — each 30-second increment is one frame in the final animation — and drawing out all of the paths that have been traveled up to this point. The blue blobs that you see represent specifically the last 30 seconds worth of running — the longer the blob, the faster a given runner was traveling. (In a few cases, you might see really long, fast-moving blue blobs; those are typically moments where somebody forgot to turn off Runkeeper as they got on a car or train).</p><p>As time marches forward in the animation, the resulting image gets more and more complex and takes longer for our server to draw. To speed up the process, I’m using the Python multiprocessing package to draw several frames of animation at once, netting a roughly 10x speedup.</p><p><strong>The Tools I Used</strong></p><ul><li>My final data set was runs from Runkeeper’s internal employees throughout our company history (up to mid-October 2016).</li><li>My visualization tools of choice was <a href="http://matplotlib.org/">matplotlib</a> in a <a href="http://jupyter.org/">Jupyter notebook</a>.</li><li>I used <a href="https://stanford.edu/~mwaskom/software/seaborn/">Seaborn</a> to tweak some of the visual stylings.</li><li>I did the map projections with the Python <a href="https://pypi.python.org/pypi/utm">UTM</a> package.</li><li>I used <a href="https://www.ffmpeg.org/">ffmpeg</a> to encode the resulting frames into a video.</li></ul><p><strong>What’s Next?</strong></p><p>I plan to set these animation tools loose on other cities and countries around the world, but there’s an obvious problem: Runkeeper is based in Boston. Barring the occasional business trip or vacation, most of our internal employee runs stick to the Boston area, too. The obvious solution is to look at runs from Runkeeper users, but we want to do so in a way that respects user privacy.</p><p>My tentative solution to this is to follow a few key rules:</p><ul><li>Use only runs from users that have marked their activities with publicly-shared maps. By default, activity maps are only set to be shared among a user’s friends. If you’re a Runkeeper user and you want to learn more about managing your sharing settings, <a href="https://support.runkeeper.com/hc/en-us/articles/201109286-How-To-Manage-Sharing-Settings">check out this article</a> from our Support team.</li><li>Trim off a few hundred meters from the beginning and end of each run. This will help obfuscate where users are coming from in the case that they begin or end their runs at a sensitive location (for instance, their home or workplace).</li><li>Select a random sample of activities over a long time period, so that we don’t pull in too many runs from one user or one period of time.</li></ul><p>What cities are you hoping to see mapped out with runs? Let me know in the comments here!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=33aefb1f6e89" width="1" height="1" alt=""><hr><p><a href="https://medium.com/runkeeper-everyone-every-run/heres-what-24-hours-of-running-in-boston-looks-like-33aefb1f6e89">Here’s What “24 hours” of Running in Boston Looks Like</a> was originally published in <a href="https://medium.com/runkeeper-everyone-every-run">ASICS Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Map of the Running Season]]></title>
            <link>https://medium.com/runkeeper-everyone-every-run/a-map-of-the-running-season-7c24242f8b2?source=rss-2d753462da94------2</link>
            <guid isPermaLink="false">https://medium.com/p/7c24242f8b2</guid>
            <category><![CDATA[data-science]]></category>
            <category><![CDATA[running]]></category>
            <category><![CDATA[tech]]></category>
            <dc:creator><![CDATA[Chris Drouin]]></dc:creator>
            <pubDate>Wed, 28 Sep 2016 17:47:01 GMT</pubDate>
            <atom:updated>2016-10-07T20:05:58.455Z</atom:updated>
            <content:encoded><![CDATA[<p>A recent request from my coworkers at ASICS spawned some data visualizations that were just too cool not to share.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*nTUfzUMmaOtEQu2CSP5Y1Q.png" /><figcaption>Weekday and weekend running patterns in the United States (2015, Runkeeper users)</figcaption></figure><p>These heat maps represent when Runkeeper users most commonly started their runs in the United States during 2015. Dark blue regions represent times when high numbers of users started runs, while dark red regions represent times when few or no users started any runs. The heat maps are broken out separately for weekday and weekend runs, as there are substantial differences in behavior between the two categories.</p><p>There are a few fascinating patterns that can be seen for our U.S. runners:</p><ul><li>Weekday morning runners pretty consistently get up around 6am to start their runs, more or less regardless of when the sun rises.</li><li>Evening runners, however, push later and later into the evening during the summer months, out to 9pm during mid summer. I suspect that this is either to beat the heat or to avoid getting blinded by the sun.</li><li>See those funny weekend spikes in March and November? Those are Sundays where the schedule is offset by an hour… because most runners haven’t adjusted to the start/end of Daylight Savings Time yet.</li><li>There are several weekdays throughout the year that look more like weekends, with later starts and early ends. These are holidays! Thanksgiving in particular stands out; mid-morning Turkey Trots make Thanksgiving one of the most popular running days of the year.</li></ul><p>But the really fascinating things that I saw came from looking into the patterns of other countries. Below you can see the same heat maps rendered with Brazilian user data:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*AywvYVx8ZEIgRf2VRk3M7A.png" /><figcaption>Weekday and weekend running patterns in Brazil (2015, Runkeeper users)</figcaption></figure><p>Where we saw a lot of daytime runners in the U.S., particularly around lunch time, this is almost entirely absent in Brazil. This pattern holds over the weekends, which suggests that the behavior is a reaction to weather and temperature conditions rather than workday norms. We also see that:</p><ul><li>Seasonal patterns are inverted in Brazil (and for other Southern Hemisphere countries). These patterns aren’t very pronounced, however, as most of Brazil remains fairly warm throughout the year.</li><li>Many Brazilian runners choose to run in the mid-morning on weekends. This pattern — favoring morning runs on the weekend — seems fairly common globally.</li></ul><p>Here are heat maps from a few more countries that caught my eye:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TBYran8VZnHFBXepschiSg.png" /><figcaption>Weekday and weekend running patterns in Sweden (2015, Runkeeper users)</figcaption></figure><p>Swedish runners limit their weekday running during the chilly Swedish wintertime. When summer comes, they take full advantage, particularly during the vacation month of July.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*TLHGYpSpWylnQ5I1Ny7zEQ.png" /><figcaption>Weekday and weekend running patterns in Japan(2015, Runkeeper users)</figcaption></figure><p>Two interesting patterns I see in Japan: one, we have a cohort of weekday runners that start at noon on the dot every day. Lunch-time runs are fairly common, but rarely have this degree of punctuality. Two, there are occasional weekends — particularly in November — where very few users go running. I’m curious to know if there is a cultural- or weather-related reason behind this behavior.</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*zGgEjuJxuNO7_7PxrMEiFw.png" /><figcaption>Weekday and weekend running patterns in the United Kingdom (2015, Runkeeper users)</figcaption></figure><p>Runners in the U.K. heavily favor evening running during the week; those who do run in the morning seem to come in three waves, at 6am, 9am, and noon. We see a lot of Saturday runs starting at 9am, which as Lee Cherry pointed out in the comments is likely due to the popular weekly parkrun 5K races. Thanks for the tip, Lee! (<em>Originally: On the Sundays, we see a large number of users who consistently start running at 9am on the dot.</em>)</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*UWEVPMyX4Ed1gy_GAeaVcw.png" /><figcaption>Weekday and weekend running patterns in the Spain (2015, Runkeeper users)</figcaption></figure><p>Spanish runners dramatically change up their habits during the warm late summer months. They avoid running in the early evening hours, and some even switch their routines to run on weekday mornings in August.</p><p>Got a country request? Leave me a note in the comments and I’ll see what I can come up with!</p><p>By request from Gabriele, here is a map generated from our Italian users:</p><figure><img alt="" src="https://cdn-images-1.medium.com/max/1024/1*i4GTeHh-gEv6v_L7JWFKmQ.png" /><figcaption>Weekday and weekend running patterns in Italy (2015, Runkeeper users)</figcaption></figure><p>Our Italian runners favor the morning even more heavily than our Spanish users during the summer months. I also seem some interesting weekday morning patterns near the start of April, May, and June — can anyone comment on what these might be? I see holidays (Easter, International Workers’ Day, and Republic Day) that might correspond with these dates.</p><p>Some methodological notes for the curious:</p><ul><li>These heat maps were generated using <a href="http://matplotlib.org/">matplotlib</a> in a <a href="http://jupyter.org/">Jupyter notebook</a>. I used <a href="https://stanford.edu/~mwaskom/software/seaborn/">Seaborn</a> to tweak some of the visual stylings.</li><li>I chose the <a href="http://colorbrewer2.org/#type=diverging&amp;scheme=RdBu&amp;n=10">RbBu ColorBrewer palette</a> as a relatively nice-looking and colorblind-friendly palette.</li><li>All start times shown are in the user’s local time at the time the activity was logged.</li></ul><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=7c24242f8b2" width="1" height="1" alt=""><hr><p><a href="https://medium.com/runkeeper-everyone-every-run/a-map-of-the-running-season-7c24242f8b2">A Map of the Running Season</a> was originally published in <a href="https://medium.com/runkeeper-everyone-every-run">ASICS Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
        <item>
            <title><![CDATA[A Global Pace Paradox]]></title>
            <link>https://medium.com/runkeeper-everyone-every-run/a-global-pace-paradox-e5e52186a38c?source=rss-2d753462da94------2</link>
            <guid isPermaLink="false">https://medium.com/p/e5e52186a38c</guid>
            <category><![CDATA[startup]]></category>
            <category><![CDATA[running]]></category>
            <category><![CDATA[data-science]]></category>
            <dc:creator><![CDATA[Chris Drouin]]></dc:creator>
            <pubDate>Mon, 11 Jan 2016 14:38:09 GMT</pubDate>
            <atom:updated>2016-01-11T14:38:09.811Z</atom:updated>
            <content:encoded><![CDATA[<h4>When running data is more than it seems at first glance…</h4><figure><img alt="" src="https://cdn-images-1.medium.com/max/602/1*i_DNPBUZ5Re7Z55ASurfiA.png" /></figure><p>When we launched the Runkeeper Global 5k challenge in December, one of the things we were curious to find out was which country (on average) has the fastest runners. If you’ve read our blog post on the <a href="http://blog.runkeeper.com/4318/global-5k-the-results-are-in/">Global 5k results</a>, you’ll already know that the answer is the Netherlands:</p><h4>Overall pace from the Global 5k</h4><ol><li>Netherlands (5:53 per km/9:28 per mi)</li><li>Belgium (5:58 per km/9:36 per mi)</li><li>United Kingdom (6:00 per km/9:40 per mi)</li><li>France (6:02 per km/9:42 per mi)</li><li>Italy (6:08 per km/9:52 per mi)</li><li>Spain (6:08 per km/9:53 per mi)</li><li>Japan (6:08 per km/9:53 per mi)</li><li>Germany (6:10 per km/9:55 per mi)</li><li>Australia (6:10 per km/9:56 per mi)</li><li>Canada (6:12 per km/9:59 per mi)</li><li>Russia (6:14 per km/10:02 per mi)</li><li>Sweden (6:14 per km/10:08 per mi)</li><li>United States (6:19 per km/10:10 per mi)</li><li>Brazil (6:31 per km/10:30 per mi)</li><li>India (8:00 per km/12:53 per mi…must be all that traffic, yikes!)</li></ol><h4>Not so fast</h4><p>This wasn’t actually the <em>first</em> answer we came up with. In our initial analysis, Italy (ranked #5 above) came in as the fastest, with a 9:25 min/mile median pace. The Netherlands was in second place, a fraction of a second slower. So: what changed?</p><p>The answer lies in something known as <a href="https://en.wikipedia.org/wiki/Simpson%27s_paradox">Simpson’s Paradox</a>. In our first pass, we grouped all participants together. Not all runners are alike, though. Male participants in the Global 5k were generally a bit faster — the median pace for men was 9:12 min/mile and the median pace for women was 10:42 min/mile.</p><p>Italian runners were on average the fastest, true… but this was because the mix of runners in Italy was different than in most other countries. Most European countries skewed slightly male (55% — 60% male), but in Italy fully 75% of participants were male. The boost from all those fast-running men translated into an speedy overall pace for Italy!</p><p>Despite having the fastest overall median pace, Italy’s male runners ran at a median pace 16 seconds slower than men in the Netherlands, and its female runners ran 32 seconds slower than Dutch women. With Dutch men faster than Italian men and Dutch women faster than Italian women, it really wouldn’t be fair to say that Italy was the fastest country.</p><p>To solve this issue and create a final, “fair” pace for each country, we took a straight average of the male and female paces. Calculated this way, the pace in the Netherlands barely changes (since they were already 53% male / 47% female), but the pace for Italian runners slows down by 27 seconds to a 9:52 min/mile.</p><p>Italy and the Netherlands weren’t the only countries affected by this statistical quirk. The US and Brazil swapped places, since Brazil had 72% male runners while the US had just 47% male runners. Both countries still stayed near the bottom of the overall country rankings, though… better luck next time!</p><img src="https://medium.com/_/stat?event=post.clientViewed&referrerSource=full_rss&postId=e5e52186a38c" width="1" height="1" alt=""><hr><p><a href="https://medium.com/runkeeper-everyone-every-run/a-global-pace-paradox-e5e52186a38c">A Global Pace Paradox</a> was originally published in <a href="https://medium.com/runkeeper-everyone-every-run">ASICS Digital</a> on Medium, where people are continuing the conversation by highlighting and responding to this story.</p>]]></content:encoded>
        </item>
    </channel>
</rss>