[hozn]

@acl 44421 wrote:

“Individual Elevation Gain”?!?!:confused::eek::eek:

crap

That was just an example of other things we can do. I left it in since apparently I stumbled upon a category where I am in first place Avg speed of rides is another easy one I might add — in addition to the environmental ones.

[hozn]

I’m cross-posting from the pointless prizes thread, in case anyone doesn’t care about prizes but might want to participate in data mining (or view what [little] I’ve done thus far).

Here is the freezing saddles data mining application running. This shouldn’t be viewed as official results, but does sync every 15 minutes and rebuilds daily (designed to mirror official leaderboard schedule). I can’t promise there aren’t bugs; if there is a discrepancy between data here and the official leaderboard, assume it is a bug in my app — and please let me know.

http://freezing.hozn.net/

Here is the code in case you’d like to play with / enhance this yourself:

https://github.com/hozn/bafs.git

The codebase is a little messy for the sake of putting something up there sooner than later, but would love to have help adding additional queries, visualizations etc. I plan to integrate wunderground API for things like sunrise/sunset/temperature (diurnal temperature model!), etc.

Also, if someone just wants to connect to the database and run queries, I can set you up with a read-only user. Let me know.

[hozn]

BTW, I applied some updates to my data mining site. I will start working on weather integration next.

http://freezing.hozn.net/

I can help with some of these other questsions — e.g. weekday/weekend distance is a pretty trivial query.

Right now here’s what weekday warrior looks like:

mysql> select
-> A.id,
-> A.name,
-> sum(distance) as total_distance
-> from daily_scores DS
-> join athletes A on A.id = DS.athlete_id
-> where DAYOFWEEK(ride_date) IN (2,3,4,5,6)
-> group by A.id, A.name
-> order by total_distance desc
-> ;
+

---

+

---

+

---

+
| id | name | total_distance |
+

---

+

---

+

---

+
| 1129136 | Rod Smith | 1337.2003955841064 |
| 75834 | Pete Beers | 1258.796175301075 |
| 182475 | Hans Lellelid | 1166.668829202652 |
| 754817 | Shawn Dread | 1102.5912912487984 |
| 300685 | Chris Randall | 1078.8127031326294 |
| 488016 | jeff lemieux | 1025.880711555481 |
| 120192 | Will Lynn | 966.9974447488785 |
| 186394 | Elizabeth M. | 907.9707537293434 |

But it’s a little closer for weekend warrior:

mysql> select
-> A.id,
-> A.name,
-> sum(distance) as total_distance
-> from daily_scores DS
-> join athletes A on A.id = DS.athlete_id
-> where DAYOFWEEK(ride_date) IN (1,7)
-> group by A.id, A.name
-> order by total_distance desc
-> ;
+

---

+

---

+

---

+
| id | name | total_distance |
+

---

+

---

+

---

+
| 300685 | Chris Randall | 658.0815916061401 |
| 1473093 | Justin Antos | 619.3079282045364 |
| 75834 | Pete Beers | 596.2577639818192 |
| 1129136 | Rod Smith | 559.4644396305084 |
| 550287 | Kevin Ulrich | 326.3741452693939 |
| 10446 | Sean Chisham | 323.4553096294403 |
| 784659 | Christopher N. | 317.126194357872 |

(of course you may wish to use % of riding instead of just how many miles ridden?)

There should also probably be an award for most sleaze rides, if that hasn’t already been proposed.

mysql> select
-> A.id,
-> A.name,
-> count(*) as rides
-> from daily_scores DS
-> join athletes A on A.id = DS.athlete_id
-> where distance < 2.0
-> group by A.id, A.name
-> order by rides desc
-> ;
+

---

+

---

+

---

+
| id | name | rides |
+

---

+

---

+

---

+
| 136274 | Mark Blacknell | 13 |
| 645556 | Kel Moore | 13 |
| 580874 | Mark Nielsen | 11 |
| 564600 | Kathy Lewis | 10 |
| 1224809 | Ken Buja | 10 |
| 1410917 | Karen Smith | 10 |
| 126281 | Shawn Gallagher | 9 |
| 754817 | Shawn Dread | 9 |

(I might have some bugs in these queries.)

[hozn]

@MRH5028 44182 wrote:

I feel guilty about this, but my commute to work is a mile. Ballston to Clarendon. I know I’m lucky to have an easy commute, but come next year my office is moving to the Reston area so that will come to an end. Some days I’m “lazy” and just walk instead of ride.

Hit me up when you move and we can coordinate commuting our together sometime. (I commute from Westover to Reston.) I actually think I am lucky to have a longer commute; it makes me get exercise I need without me having to make special time for that. And an hour is good decompress time coming home from work.

Sent from my Nexus 7 using Tapatalk HD

[hozn]

@DaveK 44129 wrote:

http://www.bikesdirect.com/products/motobecane/lechamp_cf_carbon_frameset.htm

Cool. Although anyone who wants me to buy a bike sight unseen and doesn’t provide stack and reach should be flogged. I know it can be calculated, but I’m a lazy American who doesn’t believe in science.

I love my Six13 but this is a pretty good budget option should something go wrong…

Yeah, I’ve never seen a site list stack/reach for frame geometry. For road bikes it would be nice to have more nuanced fit information. Then again, my fit seems to be constantly evolving depending on how much & what kind of riding I’m doing; maybe in 10 years I’ll have it nailed down to the mm.

The LeChampion was well reviewed so I decided it was worth the small premium over [aesthetically] similarly Chinese frames on ebay. It’s not very light for a CF; I believe my 58cm was 2lbs 12oz, but that is perfectly fine by me — it’s still a ~16.5lb build, and I suspect that is not what’s holding me back

Nashbar has an arguably better-deal carbon road frame option too: http://www.nashbar.com/bikes/Product_10053_10052_543482_-1___

(And today it would be 20% off.)

That is probably a better deal since the included fork is full carbon — and the tapered steerer is probably the way to go

My next road frame will be a disc frame. After riding carbon wheels for a week or so, I’ve learned [what everyone already knew] that a carbon braking surface is a horrible idea.

[hozn]

I have gotten my freezing saddles data mining codebase to a ready enough state to post about it. It’s very preliminary (and very ugly) and doesn’t do much of anything yet, but the API is working and the data is there and there’s lots of potential and I plan to continue development on this and am open to suggestions — and Git pull requests

Here is the code in case you’d like to play with / enhance this yourself:

https://github.com/hozn/bafs.git

There are some crude setup instructions there, but feel free to post or PM me if you have problems or more questions.

Here is the web application running. Right now it’s just showing some barcharts of team & individual leaderboards (the data may be a bit stale as at this time I have not yet set up the cron jobs to sync the data).

http://freezing.hozn.net/

I will probably turn my attention to the UI for a little bit — enjoying getting a chance to play w/ Google’s charting API — but also plan to add wunderground integration to the sync script to pull down general weather data for each ride.

Also, if anyone would like read access to the MySQL database under the hood here (to run your own queries, etc.) please PM me and I’ll setup a user / send you the creds.

[hozn]

@JimK 44090 wrote:

Thanks for the suggestions. I blew it out with air (thanks hozn). The pump is working again and I will be preemptively replacing tubes every few months (if they last that long without a flat).

This is a good cautionary tale; I’ve never heard of this happening. I’ve definitely used tubes for up to a year or so before having to replace them, … but I will probably be more mindful of this. For my tubeless setups I definitely have valve stems that have lasted for multiple years now, but not under quite as much use / frequent pumping as my daily commuter.

[hozn]

My commute is 16 miles each way. It’s just out the W&OD so it takes me around 50-55 minutes moving time on a normal day, and 55-60 minutes total travel time. I think it is a perfect length commute, since it provides a good workout without huge daily obligation (or take away from family life). I also love that I can detour through Lake Fairfax Park mtb (or cx, in my case) trails.

[hozn]

@DaveK 44102 wrote:

Did you piece it together or buy it and strip/paint the frame? The specs on those are great, but I can’t stand the paint schemes.

Oh, bikesdirect is selling these frames in matte black, gloss black and gloss white for $599 (frameset, but I just resold the fork since I had the Enve which I like very much). Agree that the standard Motobecane paint schemes are atrocious.

[hozn]

Obviously hard to say w/o seeing it, but is it loose in there — i.e. would something like a magnet help? (I’m not sure what metal those are made of, but figure it’s worth a mention.) Or blowing air through the opposite end of the pump somehow … or putting super glue on the end of the now-broken presta valve stem and jamming it in there to reclaim it’s lost member? (that last suggestion could probably backfire…)

[hozn]

I can’t speak for the MS ride, but I did the TdC in 2011 in Reston; it was a great experience. I organized a small (8-person) team for my office (which matched our $$ raised) and rode the century with a couple friends/teammates. The ride itself was great — especially once you get out past Leesburg. And the support was phenomenal — lots of very well-stocked aid stations. My only negative comment would be that on the team organization front they didn’t have a system whereby corporate donation matches could be distributed equally among the team members. But overall I definitely would recommend TdC.

[hozn]

@Dirt 43914 wrote:

Lovely bike! I like the stealth look.

Thanks! Yeah, me too. I really like the new Foundry disc road bikes/frames, but was too impatient to wait for them to become available (and then to go on sale…).

[hozn]

My new road frame/build (Motobecane LeChampion CF).

[hozn]

Hey, vvill do you want a ritchey carbon cx fork (canti brake posts) for your project? Alu steered but should be lighter than steel

[hozn]

I can also help with data mining. I’m working on a GAE (Google App Engine) app to collect the rides and make the data available to whomever wants to play around with it (CSV & GPX when available, etc.). I think it’ll probably still be a week or so before there is something to post on that.

I looked into wunderground mashup earlier; getting general historical data is easy. Getting hourly data (e.g. to actually determine how cold it was during a ride) may be trickier (may have to guess based on trends of coldest hours during the day, etc. … aka, as a quick Google just taught me, use a “diurnal temperature model”.

[Author]

Posts