Developing Training Software

timbradley said:

Hi All

I am initially going to develop it to use Polar HRM files as a basis then expand from there.

It isn't that I am particularly unsatisfied with the Polar Precision software, I just thought I'd use it as an exercise in software development as I am currently doing a Masters in IT.

Anyway - The question I want to ask is:

What features would you like to see in such a piece of Training Software?

Wether compared to Polar SW, I'd like it to include Polar features, and additionally:
-automatically removing too high readings caused by high voltage electricity lines or shirt flapping. Currently needs manual editing for replacing e.g. 235 bpm numbers.
-the energy consumption or calory calculation needs minor tuning, e.g. if I keep the HRM at the work, sitting most of the day, it shows 0 cal usage for the whole 8 hours

Antti

a handy feature would be one that you could input you Kcal consumption and have it automatically compare this to the estimateded expenditure.

sidewind said:

Wether compared to Polar SW, I'd like it to include Polar features, and additionally:
-automatically removing too high readings caused by high voltage electricity lines or shirt flapping. Currently needs manual editing for replacing e.g. 235 bpm numbers.
-the energy consumption or calory calculation needs minor tuning, e.g. if I keep the HRM at the work, sitting most of the day, it shows 0 cal usage for the whole 8 hours

Antti

Thanks for the input - Will look at these. Some initial thoughts on these:

- The high reading removal can be easily done, I guess the only tricky thing there will be deciding what is a high reading for an individual - I guess an estimated Max HR might be the best way to limit readings - will think about it (Might have to break out the old university statistics textbooks!!)

- Fine tuning the energy consumption will be a little more complicated - the elevated heartrate is the basis for the calculation, as far as I'm aware - I suppose if the activity were selected for any 'sedentary' time (sleep, office work etc.) published values could be used...I'll try to incorporate it that way!

jcjordan said:

a handy feature would be one that you could input you Kcal consumption and have it automatically compare this to the estimateded expenditure.

-Should be easy enough!

Keep the ideas coming!!

I would like to generate a report graph that compares values over two different time ranges... e.g. Distance, avg speed and avg HR for Jan 2005 compared to Jan 2006 (on one page).

Would like to be able to change the font size in the reports.

Would like to be able to group specific rides and then run reports on only those (in addition to the current capabilities with all rides). I would find this useful for comparing HR, speed, time, etc for routes that I use as tests of my fitness that I do every few weeks.

timbradley said:

Thanks for the input - Will look at these. Some initial thoughts on these:

- The high reading removal can be easily done, I guess the only tricky thing there will be deciding what is a high reading for an individual - I guess an estimated Max HR might be the best way to limit readings - will think about it (Might have to break out the old university statistics textbooks!!)

- Fine tuning the energy consumption will be a little more complicated - the elevated heartrate is the basis for the calculation, as far as I'm aware - I suppose if the activity were selected for any 'sedentary' time (sleep, office work etc.) published values could be used...I'll try to incorporate it that way!

-Should be easy enough!

Keep the ideas coming!!

I was just thinking on the lines of having a section where you could input your values for Carb, Pro, Fact and KCal's and let you compare this to the estimated expendeture that comes with the polars

It should include basic TRIMP calculation I guess.

an easy way to compare performance gains over a certain amount of time (easy to read graphs), but maybe they already have that

Falstaff said:

I would like to generate a report graph that compares values over two different time ranges... e.g. Distance, avg speed and avg HR for Jan 2005 compared to Jan 2006 (on one page).

Would like to be able to change the font size in the reports.

Would like to be able to group specific rides and then run reports on only those (in addition to the current capabilities with all rides). I would find this useful for comparing HR, speed, time, etc for routes that I use as tests of my fitness that I do every few weeks.

I accomplish this by giving the specifc route a sport name unto itself, then I include only that "sport" on the report. It lets you filter out the other routes, recovery rides, intervals, etc. so that you're comparing your progres on that route only. I realize it's not the "best" way to accomplish it, but it works till there is a better solution.

Polar's reporting features are pretty good, but some thoughts:

1. Currently, you can only average one measurement over a time interval. Say...measuring average speed by month. You have to choose the initial bar graph(doesn't have to be a bar though) to do this, and then tell it to total by day/week/month, etc. You can add other measurements onto this same report, but you can't have it average over the period, you get a data point for each exercise during that time period...which gets a little too "busy". You can sort of get around this by using trend lines....but I'd still rather be able to have full funcitionality with each measurement included on the graph.

2. Would be nice to have the ability of deleting a certain type data from an exercise. Take power for example, I had a malfunction on a couple of rides that gave very erroneous power values. I could edit all of them to something I "guessed" at, but I would rather just remove the power measurement from that specific exercise...while keeping the data for the other exercises. It does not allow you to do that. If you set it all to 0, then it reports 0 as opposed to "null".

3. The current software will let you see your averages either with, or without the 0's averaged in on any particular exercise. I'd love to have the option of turning "0 averaging" on or off as a global default.

4. If you're doing a "calendar view", it would be nice to have the ability to be a little more descriptive on the exercise name. It currently will include the abbreviation as you have defined it, but I'd almost rather see the exercise or "sport" name as the top line of a particular day. I suppose if you have more than one exercise on a given day, it could list multiple in a list.

Just food for thought,

John

timbradley said:

Hi All

I am currently thinking about developing some a new piece of Cycling specific training software.

I am initially going to develop it to use Polar HRM files as a basis then expand from there.

It isn't that I am particularly unsatisfied with the Polar Precision software, I just thought I'd use it as an exercise in software development as I am currently doing a Masters in IT.

Anyway - The question I want to ask is:

What features would you like to see in such a piece of Training Software?

All answers will be appreciated and considered - If this grows enough, I may consider using it as the basis for my degree project! I will also keep you all posted on progress!

Look forward to hearing from you all!!

A first nice thing could be for example a (Java?) library to manipulate polar files. Very nice would also be the same for the other formats (hac4, powertap, srm, etc...), so that you can use data from any of these formats (they represent more or less the same kind of data).
Such a thing would be a great base for a lot of training programs. In my opinion libraries available at the moment are quite crappy.

I believe building modular components is better than a big monolitic block, so for example another possible thing could be a library that draws graphs from these files (using the libraries mentioned before).

Just some ideas...

frenk said:

I believe building modular components is better than a big monolitic block, so for example another possible thing could be a library that draws graphs from these files (using the libraries mentioned before).

I agree with your Ideas - and will try to keep the program as modular as possible to allow for future upgrade/expansion.

I am initially developing the program based on the HRM file format, but will endeavour to expand it to utilise other formats after I have developed it to a reasonable standard.

BTW, I will likely be using VB.NET

Regarding the intial thoughts on sedentary hrm data, some of my clients appreciate being able to distinguish between baseline rates (calorie expenditure,etc) and the elevated ones due to the exercise, so software that could sort out and archive both types of data would be useful (though, of course, the sedentary data is based on earlier readings or estimates).
J. Binkovitz, CSCS
www.binksgym.com

- Fine tuning the energy consumption will be a little more complicated - the elevated heartrate is the basis for the calculation, as far as I'm aware - I suppose if the activity were selected for any 'sedentary' time (sleep, office work etc.) published values could be used...I'll try to incorporate it that way!

-Should be easy enough!

Keep the ideas coming!![/QUOTE]

the most amazing thing would be something similar to TSS score but takes into account weather, altitude, diet, and previous days training. If something like this could be formulated that would really take things to the next level.

timbradley said:

Thanks for the input - Will look at these. Some initial thoughts on these:

- The high reading removal can be easily done, I guess the only tricky thing there will be deciding what is a high reading for an individual - I guess an estimated Max HR might be the best way to limit readings - will think about it (Might have to break out the old university statistics textbooks!!)

- Fine tuning the energy consumption will be a little more complicated - the elevated heartrate is the basis for the calculation, as far as I'm aware - I suppose if the activity were selected for any 'sedentary' time (sleep, office work etc.) published values could be used...I'll try to incorporate it that way!

-Should be easy enough!

Keep the ideas coming!!

Are you making a commercial program or some sort of freeware or shareware?
Since this is for your masters does that mean you can't work with anyone else?
I've already made a little C program that reads PowerTap .csv files and *fixes* hr data, plus calculateds NP, TSS and IF. For me there were two issues with bad heart rate data - zero values in the middle of non zero values and values more than 10 over my max of 202 or whatever both of which I replaced with a smoothed value. I think some folks would rather have the bad data in there so they know there was a data drop but to each his/her own. I stopped working on it a couple of years ago, but should probably just finish it.

the holster said:

previous days training. If something like this could be formulated that would really take things to the next level.

It's already being worked on.

Woofer said:

Are you making a commercial program or some sort of freeware or shareware?
Since this is for your masters does that mean you can't work with anyone else?
I've already made a little C program that reads PowerTap .csv files and *fixes* hr data, plus calculateds NP, TSS and IF. For me there were two issues with bad heart rate data - zero values in the middle of non zero values and values more than 10 over my max of 202 or whatever both of which I replaced with a smoothed value. I think some folks would rather have the bad data in there so they know there was a data drop but to each his/her own. I stopped working on it a couple of years ago, but should probably just finish it.

It's not actually for my masters - Just a side project - Wasn't really thinking commercial at this stage - just something that I, and hopefully others out there, would use and find helpful...but you never know what the future holds!!

Also - just as an aside - am thinking I may actually write the program in Java rather than VB.NET - mainly because it is more useful for the purpose of my masters

Yeah stick to something cross-platform for all of us Linux and Mac users (I see a penguin in your logo, are you a linux user perhaps??).

Good luck with this project, make sure you post to this forum when its done!

daniels said:

Yeah stick to something cross-platform

Yeah - OK - will stick to Java as the development language.

However, having a limited knowledge of Java (Although I am working on it) will slow progress a little - at least at the start!

Thanks for all the suggestions so far! Keep 'em coming - Whilst I obviously won't be able to include them all, the bigger the list to choose from the better!