Club API Feedback and Improvements

[Winterspite]

Hi, I'm trying to build a small app for my company's Strava club to track weekly "winners" similar to how the club page shows "Distance / Total Running Time / Climbing" leaderboard entries and ran into a few issues in the process that I'd like to raise as well as a suggestion on how to mitigate why I think you built the API the way you did.

My goal is to track the 1st/2nd/3rd place for each week and accumulate "medals" for people over the course of the year. Additionally, I'd like to store and keep certain things updated such as "Longest Run This Year" or "Most Elevation In A Run This Year" and post when someone sets a new record. However, I ran into a few issues with the Clubs API and would like to see if there's a different way to accomplish this OR if I can make suggestions instead:

• Club API doesn't return a timestamp for activities when you call getClubActivitiesById. That's fine, I can work around that by just calling via cron job every 15 minutes and can roughly figure out when new activities are added. Ideally I could subscribe to a webhook for notification on new runs, but that one isn't really a big deal. I will note that the API doesn't explicitly say if the results are in any kind of a sorted order (although they appear to be from my testing).
• Club API doesn't return a unique activity ID when you call getClubActivitiesById. This one isn't great. I can try to generate my own unique hash by grabbing hopefully deterministic fields like moving time, distance, etc. - but I also know that people can go edit their runs to trim them, so that might change and break my way of knowing "have I seen this run / added it to my tracker yet or not?"
• Club API doesn't return a unique athlete ID when you call getClubActivitiesById. This is unfortunately the thing that killed my project. All I can see when making the API call is First Name + Last Initial. Even though I can see the full names of all of the users in my web browser, the API won't tell me more. Also, this assumes that a user never changes their name, because then my tracking will break.

API response

I understand for privacy reasons, you may not want to return the exact activity ID or athlete ID that you use elsewhere, but perhaps as a workaround you generate a per-club GUID? I don't need to know the athlete's global Strava ID, I just need to know in a consistent and deterministic manner which "Joe B." posted a run so I can accurately track leaderboards. If there's a different API endpoint I should be calling instead, or something I can use to gather additional details about the activities and athletes, that would be great as well.

My only other thought is to instead just have all of the folks from my company's running club individually authorize my app, at which point I can scrape their activities as individuals and generate my own leaderboard from the people who authorized the app, but then that completely defeats the purpose of having a Club, as we won't be relying on that at all.

Thanks!

[antban]

Hi, I am trying to understand now if there are any updates/real solutions here or is it just forgotten?

Also I wanted to remind that on the API page ClubActivity (https://developers.strava.com/docs/reference/#api-models-ClubActivity) still contains MetaAthlete, while in fact returns ClubAthlete. Can you please fix documentation not to have misleading information (and add/mention "before"/"after" parameters that are either supported or not supported).  

I am also trying to understand is there anything but web scraping left for developers? (I am talking about public activities)

[antban]

Just trying to clarify myself: 

1. Scrapping is not legal according to https://www.strava.com/legal/api(to anyone who did not know that)

2. There are solutions that are analyzing different secondary parameters to match activities.

So official way allows to achieve results, but requires a lot of efforts and is very inefficient (but still doable), non-official way is not legal, requests to get some form of solution are ignored.

Can someone please direct me to a place where privacy policy was discussed and why public activities are violating privacy. 

[imadeous]

1. Why has this issue not been resolved despite the fact that every single developer is facing it and are requesting Strava to make a workable solution instead of the developers coming up with other means?

[tpreece]

I'm having the exact same stuggles trying to track activities of a cycling club. Having to do all sorts of hacks to establish a date, the athlete and trying to prevent duplicates. Seems adding a date and a uniqueID would be simple API adds.

[petrepa]

We are developing a dashboard which will be situated in the entrance of our office. We planned to have a summary of our company Strava club (how much we have run this week and this month and so on). This could have been a motivating feature, but because of the lack of timestamps in the activities this functionality is not possible to obtain in the current state of the API. Hope this can be fixed somehow in the near future.

[bobbaldhead]

Basically having the exact same requirements ... i found out there is a parameter "after" accepting a unix timestamp which i could not find in the docs (at least not for the club activities endpoint), but i seems to work (just some quick tests). Might help at least avoiding the cron job.

[SkippyVasquez · Accepted solution]

@Winterspite and @dblock, thank you for your posts. We are always looking for ways to improve the API and have made a note of this internally.

[imadeous]

Don't just mark unresolved issues as 'solutions' and sweep them under the rug please. We need real solutions.

[dblock]

You are right not to rely on club activities. They are only useful as a way to display something minimal. The additional problem is that even when the user connects their account, clubs return an anonymous version of the activity. This causes us to have to de-dup activities by comparing a bunch of fields. I think your GUID idea for an athlete ID is interesting, but it would effectively become a second, public ID for a user in the club, and for public club information would make it one little step easier to de-anonymize that "Joe B.".

I think Strava removed all this information to prevent people from building competitive club apps, not to really ensure their privacy. Which is a sad strategy. If someone at Strava is reading I'd return full activities in the club feed if the user authorized my app, and, if the argument is really not a privacy one, only for Premium users.