2022 FRC Robot Data is Beautiful! See if you can find your team!

Hot off the press, here is the link to visualize Caleb Sykes’ Scouting Database. Currently, it contains week 1 data. The plan is to update this viz every week once Caleb’s Scouting Database is updated after every round of events.

This is a handy tool for pre-scouting your upcoming events because you can see how teams performed at prior events. Plus it’s fun to see where your robot ranked alongside the others at an event, in your state/province/country or across all FRC teams.

As we all know, teams will continue to improve throughout the season and how a team performed at one event doesn’t necessarily predict how they will perform at their next event, but their past performances are a good indicator.

The viz doesn’t have it yet but I’ll be adding a dropdown for upcoming events that will display all the teams attending the upcoming event if they have been to at least one prior event. That way, you can see all teams displayed together with just a click.

Congrats to 1678, 2910 and 1690 for setting the bar super high for the rest of the FRC world. You three knocked it out of the park week 1! Awesome job!!

57 Likes

Any chance for team # labels to be added? I found it hard to figure out whos’s who otherwise without tons of hovering, which isint ideal.

Looks great otherwise. Always fun to look at how much 2910 dwarfs the rest of the district…

Edit: updated the image to one with labels

14 Likes

Done!

7 Likes

I really love this, is there anyway to establish rankings per category?!
As in what position a team is in for average auto, tele, endgame, etc…

2 Likes

Great suggestion. There is a different view I created, a stacked bar-chart view that may help with what you are looking for. Let me add some filters to it and publish. Will let you know when it’s available.

2 Likes

Here is the “Ranking” chart. Please let me know if you have any questions!

2 Likes

seems like the climbs arent being calculated correctly? 7461 was definitely without a high climber, and this adds to over 120%…

2 Likes

That is odd! I’ll take a look at the data now. Thanks, @MikLast !

1 Like

The calculation has been fixed. There is a .05 in the Traversal Rung climb for 7461… could they have been awarded a Traversal Rung climb because of a penalty? If so, we can’t differentiate between that and a valid Traversal Rung climb because we are pulling the data from TBA’s API.

1 Like

7461 was credited a traversal climb in qf 3-2 because of contact by 1778. You can see it on youtube here: https://youtu.be/ZHaml2l1WHM?t=138

3 Likes

If you’re looking to pick an alliance partner that can contribute endgame points, you’ll have to actually scout. Calculated contribution will be unreliable.

3 Likes

I love this! I’m continuously impressed by the way the FRC community can manipulate data!

Amazing
Beautiful visualization

We could’ve been pretty close had we not had power issues in on day 1

1 Like

This is my favorite thread every year. I love bubbles! :smiley:

4 Likes

There are some issues with the data from the Waco event (possibly others). I believe one of the ball sensors was not working properly and data had to be counted/entered by hand. Most of matches prior to Qual 59 have a 0/0 for auto and teleop cargo count on TBA but there are cargo points. Guessing your algorithm uses cargo count to differentiate between high/low balls?

Good question, @Brian_Selle . @Caleb_Sykes and I are having a zoom meeting this evening re: the Scouting Database so we’ll make sure to discuss this point and let you know the outcome. The ol’ data reporting caveat, “bad data in, bad data out” might be at play here but we’ll see if we can code around it.

Thanks for helping run the analysis through Quality Assurance! :wink:

2 Likes

I believe AliciaThe following data points in my scouting database, all of which will be incorrect because of the data quality issues at Waco:
auto Cargo Lower
auto Cargo Upper
teleop Cargo Lower
teleop Cargo Upper

The total auto or teleop points fields are correct in my database, but if Alicia uses those values in place of summing lower and upper scores, there will appear to be internal inconsistencies for Waco teams. I don’t know which is the better option here. .

Thanks @Caleb_Sykes. I can create a function in Tableau that can pull in the Total Auto and Total Teleop points when they have values but the upper/lower cargo for both are 0. We’ll miss the granularity that comes with knowing the upper vs. lower but at least we’ll have the total instead of 0. @Brian_Selle, I’ll get that update in this weekend.

1 Like

It isn’t just cases where upper and lower are zero that is an issue. If you look at 148 at Waco, they have non-zero values for upper and lower auto and teleop counts:

But their total auto calculated contribution was 15 and their total teleop cargo contribution was 26. The problem is that some of their matches report 0s for the counts while reporting a non-zero cargo score. To filter, you would have to calculate if the score implied by the upper and lower counts match the actual reported score.

Any qual match listed here will cause a discrepancy for every team at the event between the sum of their upper and lower calculated contributions and the score calculated contributions. Those events are:
bcvi
caph
inkok
isde1
midet
mifor
miket
mndu
mndu2
nhgrs
njbri
tant
tuis3
txwac
va305

Which is most of them unfortunately :frowning: