This year my team has taken a big step forward. We now have six netbooks that we distribute to scouts in the stands with a scouting form on each one as an excel spreadsheet. The scouts create a new spreadsheet for each team, saving it as one workbook.
In total, we want it to run like this:
Scout make workbook of scouting form spreadsheets -> Upload to FTP server laptop -> Server takes separate excel workbooks and compiles all the spreadsheets in each workbook into one single workbook -> Server laptop runs a macro to give us our rankings
We still have a few hurdles to overcome before our next district competition though, as the netbooks we use are school property, and therefore do not allow you to run batch scripts to do the FTP ( :mad: ). Also, we do not have a way or auto-compiling multiple workbooks into one workbook at the moment.
Right now, were OK with having the scout manually run a program to upload the file to the FTP server, so thats not an issue.
Do you guys have any suggestions on how we could do these two things auto-magically?
Also, (and more importantly ) how does your team do scouting?
In the past, we’ve always had clipboards, 6 at a time, for scouting robots, with sheets being returned to a central data entering hub. This year, we were lucky enough to borrow 20 iPads from our school, so we wrote a scouting WebApp. If wifi is up in the arena, we upload data as soon as the match ends from the iPads to the database.
We have 6 Nintendo DSs with custom programs. We then feed that data - I don’t have the specifics on how - into excel, which is then analyzed by another program (which can do a lot of alliance selection stuff as well.)
Our team is “old school”.
We use a single sheet for each team.
With a photo of the Robot.
Each student follows one robot for a match.
We organize the sheets in an accordion folder. We use seven students.
One for each team in a match, and one who retrieves and passes out the sheets.
We gather on Friday night to share information and opinions. We especially begin discussing who we would pick as a #1 seed, and as a #8 seed.
We discuss “dark horses”.
We rank teams according to our criteria.
Saturday proceeds and we stay in a group and adjust our list.
The use of iPads, Wifi, NetBooks, Excel, etc… is a ruse, IMHO.
Much more effort than is needed, with the flawed assumption that quantitative is better than qualitative analysis.
The human brain is a more effective “computer”.
Remember: There are three kinds of lies:
Dam*n Lies, &
I would agree in part. Qualitative and quantitative analysis both have their places in scouting. It is easy to get caught up in the ‘coolness’ of a team when looking at it qualitatively. Likewise, it is easy to look at quantitative analysis and choose a robot that may be strong in the point arena but unable to work effectively with an elimination alliance.
We are using both methods. Our scouts will be using papers that ask them both for qualitative and quantitative data. Our data-entry person will be entering that information into a Microsoft Access DB. Before each match that we play, we will print up a sheet that provides all of the qualitative and quantitative commentary for our drivers to make the best decisions when it comes to working with our alliance and planning for the coopertition bridge.
Now, if I could get the Access DB to work on Ipads. . . .
hey my team also goes old school but mostly b/c we cant afford the good stuff. can i see your scouting sheets to see what you look for to see if i have missed anything on making ours and maybe i can even give you some ideas.
In the past 294 has used a simple computer program (a field for team # and checkboxes for scoring & a few basic abilities) that uploaded to a text document unique to each computer. This was extremely difficult to compile, given that there was no way to automatically analyze the data or even to sort by team.
This year, one of our newest programmers created an HTML/PHP-based scouting app that automatically ranks teams based on scoring, defense, balancing, and match attendance (I think I’m missing a few variables . . . ). There are also several comment fields for qualitative analysis, although that’s secondary. The picklist can be automatically generated, but at the Los Angeles regional we revised it based on the dossiers we had for each team.
Aside from some technical glitches :o it’s worked pretty well, and hopefully in Spokane it’ll be fairly smooth.
We have redone our scouting AGAIN since I made this thread!
We now use a Java application to upload to a server w/ all of the net books. All scouting rankings are instantly available and can be seen at any time through the competition, and can also be sorted by characteristics that we want.
We use laptops feeding a custom database instead of paper, but otherwise we likely use very similar methods. I disagree that our method has any more effort than yours; indeed I believe that we exert far less effort to collect our data.
The only paper we ultimately generate is what the student takes on the field when selecting our alliance
Same here, the iPads we use actually streamline everything a ton. Before, someone had to manually enter 1 sheet of paper per robot per match into our scouting database, and even with quick typists it was a tedious and error prone job. Now, as soon as the match ends, data enters the database immediately, where it can be viewed right away by scouts, drivers, and anyone else who might need it.
If you look in the Google Play (Android) store you’ll find a MORT scouting app. MORT’s programmers wrote it and released it. They did a great job.
What we don’t release (and I’ve been joking should come with a ‘modest’ fee) is PC application that collects data from that app and aggregates it.
That’s how our team scouted this year.
We did port this to iOS but unfortunately there’s a missing aspect of BlueTooth for the iPhone and considering the number of SQL databases and compatibility issues they didn’t have time to write that sort of thing.
In the past, we have used custom-written programs to enter code. This year, we are simply using excel and a python script (written by Kristian Calhoun, who you may have seen around here on CD). We don’t do anything really whacky with the data, it just helps us compile all the numbers into a format that’s more easily readable. We don’t have any sort of complex equations that tell us who the 1 robot is. It simply provides us with an easier way of pulling up a robot’s overall stats for the event. We can view results by match or by team.
For example, if I wanted to see what 1676 did at Mt. Olive, I could get all that information in a few keystrokes. I could tell you how many balls they score in auto and teleop and if they balance on the bridges, etc. There are usually also people like me and Kristian who simply sit in the stands and watch matches and make observations. We take note on who we think is really standing out and use that to aid our decisions as well. We don’t just use the numbers. We might notice that a really good robot did poorly in one match. In that case, we find the scout who was watching them and ask what happened so we know whether it’s a serious problem (i.e. they made changes to their robot or code that affected their performance) or if it was just an off match for them (maybe they had connectivity issues or something).
We have done paper scouting and it’s a bit, well, clumsy given the type of technology we have available to us now. Combine this with your observations and you can usually end up with pretty good results.
As for compiling a pick list, we make a list numbered 1-23 on Saturday/Friday night (Saturday for districts, Friday for other events). Usually the top 3-4 spots are easy picks. The rest is where we need to do a lot more legwork. This involves using our collected data, our observations, as well as match video. It’s also good to have pictures of robots to jog your memory.