Since moving to a new office where my commute is a bit longer I have been obsessively recording my cycle to work hoping to see a gradual improvement. (N.B. I've stored them on Strava but you need to be a member to see ought but a summary page.) But it looks like any increasing fitness is not the dominant factor, wind is. We know that wind resistance has an enormous effect but is there a way to factor it in to my data? How might I measure, record, or just note wind speed through my ride and somehow factor it out when calculating performance each day?
As you say, wind influences this, as does countless other factors - traffic lights, weather (wind, wet roads, snow and ice, etc..)
However, unless you are remarkably lucky or remarkably unlucky, these variables will largely average out over time.. E.g a head wind on the way will become a tail wind on the return journey (or a windy day followed by a calm day, or the prevailing winds changing throughout the year)
It's also worth remembering that elapsed time wont just be skewed by negative factors, but also positive things like your bike-handling skills, confidence cornering, confidence at drafting other vehicles etc
The main point is: You should not judge fitness improvement based on specific rides ("todays ride took 2 minutes longer than yesterday") but instead look at trends over time. For example, here is a plot of average speeds for all rides during my first year of "proper" cycling:
Day-to-day it was impossible to see any improvement in speed, it was hugely influenced by factors like the terrain - but, over time there is a clear trend.
Power output as measured by a device like a Quarq, PowerTap or SRM power meter is far more representative.
Aside from being prone to external influences, increases in speed become exponentially harder. Going from 10km/h avg to 20km/h is far easier than going from 20km/h avg to 30km/h (and then 30 to 40 is even harder)
Part of this is because air resistance approximately increases with the cube of your speed (
I would speculate that the other part is, increasing your power-output becomes exponentially harder the closer your are to your "genetic limit" (e.g increasing your 1-hour sustained power from 100w to 200w is much easier than from 200w to 300w, increasing from 300w to 400w requires far more training)
What does all that rambling nonsense about watts and exponents mean?
If you look at my average-speed plot, there's a large increase in average over the first few months (the easy 10km/h->20km/h increase), then the speed appears to plateau, despite the fact I was increasing in fitness over that time...
This is where power meters become especially useful: It more directly measures a useful metric of fitness, essentially how hard you can pedals for a given duration.
I bought a power meter near the beginning of Dec 2012, where the avg-speed line is quite flat. I knew I increased in fitness since then, however the average speed line doesn't really reflect this.
The best-efforts power curve far better reflects my improvement than avg-speed:
As you can see, despite no large increase in avg speed, my best effort 1-hour power increased from 197w to 254w, which is a substantial change (diving by 60kg and using the easy-to-misinterpret "watts per kg chart", it's an increase from "moderate" to "very good")
Power meters produce great measurements, but, they are expensive.
iBike sell several versions of a bike computer which measures wind speed.
It's sold for two reasons, the first is to estimate power output without a power meter (the iBike Newton is quite expensive, but cheaper than a direct-force power meter like Quarq, PowerTap etc)
The second reason is if you have a power meter, it can use the power output, ground speed and wind speed to estimate your aerodynamic drag (since your speed on flat terrain is primarily based on power output divided by aerodynamic drag)
In short no.
You're actually better off measuring power, through a powertap, SRM or similar ergometer. This gives you a measure of your physical performance that is absolute. So your speed is a factor of your power output, chain efficiency, wind resistance, drag, tyre performance, road surface, atmospheric pressure, and gradient. It is almost impossible to measure all of these individual factors, though we have good approximations.
The power that you produce is the base of all these equations, so measuring the influencing factors is not only hard, but it will give you an inaccurate picture of your physical performance.
iBike is a power meter that uses the same thinking as yours. Instead of measuring forces produces by your legs, it calculates power based on speed, gradient and wind resistance. So what you get is power measurements that has wind resistance factored in as you want, plus gravity and other resistance. see http://ibikesports.com/how_iBike_works.html
Compared to other power meters, it looks like a pretty good cheaper alternative, it's still not cheap.
Depending on where you live, wind can be highly unpredictable and fickle. I use Google weather (google for "wind speed" and then click on "wind"; n.b. this apparently won't work in some mobile browsers) to get a rough idea of what the winds are going to be like for the day. However, I have found that even this can be horribly inaccurate if you ride any place hilly. In my experience the wind direction can shift 90 degrees between valleys and the ridges, and can even turn 180 degrees along the course of the ride, meaning I start with a nice tail wind, and then half way through the ride I run into a head wind.
Since you are commuting, I wouldn't count on a head wind in one direction to imply a tail wind in the other. This will depend a lot on where you are, but in my experience it is often not true.
When I ride I tend to look for flags along the way to determine the direction and rough speed of the wind. This allows me to gauge roughly what the wind is like, and if you want to you can make a note of these along the way and write them down in a log when you get a chance. The downside I have found with this is that often I will start depending on a flag in a particular place and then suddenly it will get taken down.
An anemometer mounted on the bike has been mentioned, but one issue with this is that it doesn't consider your riding profile. If you are riding in an aero position and then switch to standing to climb a hill, even with the same wind speed and direction you will encounter different wind forces. I don't think subtracting off your speed would be a big deal, but you would need some kind of sensor to try to determine where you were sitting, standing, hunched over, etc. and this could be tricky and error prone or at the least tedious.
Since you are using Strava, one thing you can do to try to factor out the wind in determining your fitness and performance level is to take advantage of Strava segments. Break your ride down into smaller pieces, using your knowledge of having ridden the route before to decide what the optimal segments are. This may already be done for you in places. If you are concerned about having lots of segments you can make them private so it only compares your performance to yourself. Then, the idea is to concentrate your efforts on the segment(s) that have the most favorable wind for the day. You won't be able to compare your entire route every day, but depending on conditions you might not want to push yourself hard every single day anyway.