Upvote Upvoted 11 Downvote Downvoted
TF2: 2018 Summer Data Project
posted in Projects
1
#1
0 Frags +

Something we have always noticed and been interested in about the Team Fortress 2 competitive community is the number of people that have suffered from, or currently suffer from, depression. A note regarding this topic is that many people feel depressed when they have not been clinically diagnosed, and we will not undermine your feelings, but we would like to make the distinction. Depression can be difficult to talk about, and though you may not have clinical depression your feelings still matter and we would like to note any correlations found specifically between competitive TF2 and depression. With this regard, other general demographic information will be requested in this survey to help us compare the two main ideas of the study, competitive TF2 and depression. This survey has been formatted for players all around the world, but for 6v6 and Highlander players only as these are the two main game modes for competitive TF2. One last element in the survey is the idea of how a player thinks of themselves in regards to skill level. This idea is hard to measure in survey format, but the questions were formatted in such a way so that very little information is assumed in not only answering the questions, but examining the answers to the questions as well. This project was started as a way to learn and practice elements of data science. This survey will be ran for 2 weeks. After the survey ends, we will run an analysis of the data using R. When that is done we will post a link to a website with the results and analysis.

It is very important for us that you try your best to answer these questions with complete and total honesty and thorough attention so that the data can be as clean as possible for the analysis portion of the study. Your participation in this survey is greatly appreciated. All respondents will remain anonymous throughout the entire survey process, and no questions that may reveal exactly who you are as an individual will be asked. Full participation of this survey will require you to answer at most 13 quick questions. Depending on how you answer, or don’t answer, you may have a lot less. Thank you.

https://esurv.org/?s=MOHKGG_ad48688b

Something we have always noticed and been interested in about the Team Fortress 2 competitive community is the number of people that have suffered from, or currently suffer from, depression. A note regarding this topic is that many people feel depressed when they have not been clinically diagnosed, and we will not undermine your feelings, but we would like to make the distinction. Depression can be difficult to talk about, and though you may not have clinical depression your feelings still matter and we would like to note any correlations found specifically between competitive TF2 and depression. With this regard, other general demographic information will be requested in this survey to help us compare the two main ideas of the study, competitive TF2 and depression. This survey has been formatted for players all around the world, but for 6v6 and Highlander players only as these are the two main game modes for competitive TF2. One last element in the survey is the idea of how a player thinks of themselves in regards to skill level. This idea is hard to measure in survey format, but the questions were formatted in such a way so that very little information is assumed in not only answering the questions, but examining the answers to the questions as well. This project was started as a way to learn and practice elements of data science. This survey will be ran for 2 weeks. After the survey ends, we will run an analysis of the data using R. When that is done we will post a link to a website with the results and analysis.

It is very important for us that you try your best to answer these questions with complete and total honesty and thorough attention so that the data can be as clean as possible for the analysis portion of the study. Your participation in this survey is greatly appreciated. All respondents will remain anonymous throughout the entire survey process, and no questions that may reveal exactly who you are as an individual will be asked. Full participation of this survey will require you to answer at most 13 quick questions. Depending on how you answer, or don’t answer, you may have a lot less. Thank you.

https://esurv.org/?s=MOHKGG_ad48688b
2
#2
12 Frags +

When you asked if you are unemployed, it is probably good to put in full time education as well

When you asked if you are unemployed, it is probably good to put in full time education as well
3
#3
19 Frags +

u better not be selling this data

u better not be selling this data
4
#4
0 Frags +

Bump. But while I am here I just want to restate that our goal was to only look at very surface level stuff, obviously given the basic questions, and not find a lot of deep information regarding tf2 and depression among other things. Trying to practice R skills ;). Thanks.

Bump. But while I am here I just want to restate that our goal was to only look at very surface level stuff, obviously given the basic questions, and not find a lot of deep information regarding tf2 and depression among other things. Trying to practice R skills ;). Thanks.
5
#5
3 Frags +

Who are u

Who are u
6
#6
2 Frags +

SO I FELL DOWN SOME STAIRS

SO I FELL DOWN SOME STAIRS
7
#7
-6 Frags +

this looks like one of those shitty surveys you get for "free stuff" that's a obvious scam/time waste

this looks like one of those shitty surveys you get for "free stuff" that's a obvious scam/time waste
8
#8
0 Frags +

Pretty annoying filling this out on my phone because the next page button is right next to a bunch of ads so I would accidentally clicked on then instead (pretty sure Google surveys don't have ads and there's better platforms, for the future)

Pretty annoying filling this out on my phone because the next page button is right next to a bunch of ads so I would accidentally clicked on then instead (pretty sure Google surveys don't have ads and there's better platforms, for the future)
9
#9
0 Frags +

final bump before end of survey

final bump before end of survey
10
#10
2 Frags +

what are your hypotheses? How do you intend to analyze this data? Regarding tests etc

what are your hypotheses? How do you intend to analyze this data? Regarding tests etc
11
#11
-1 Frags +
aratingawhat are your hypotheses? How do you intend to analyze this data? Regarding tests etc

u depressed m8

[quote=aratinga]what are your hypotheses? How do you intend to analyze this data? Regarding tests etc[/quote]
u depressed m8
12
#12
1 Frags +

nah i'm just another R enthusiast

nah i'm just another R enthusiast
13
#13
1 Frags +
aratingaHow do you intend to analyze this data?

I have not thought too much about that yet. I plan to run a couple regressions and make some graphs/charts that display the results of the data. I plan to graph one variable up against one or more, or at least the variables that were deemed significant by the regressions. I am somewhat of a beginner R user and I am doing this as a way to practice it with a scenario made by me from scratch. In terms of hypothesis I will talk more about that in the final product, but the basic reason I wanted to do this is to look at a correlation between tf2, depression, and other variables. If you have any suggestions or ideas I am glad to hear them.

[quote=aratinga]How do you intend to analyze this data?[/quote] I have not thought too much about that yet. I plan to run a couple regressions and make some graphs/charts that display the results of the data. I plan to graph one variable up against one or more, or at least the variables that were deemed significant by the regressions. I am somewhat of a beginner R user and I am doing this as a way to practice it with a scenario made by me from scratch. In terms of hypothesis I will talk more about that in the final product, but the basic reason I wanted to do this is to look at a correlation between tf2, depression, and other variables. If you have any suggestions or ideas I am glad to hear them.
14
#14
5 Frags +

well, my honest opinion
I don't think you'll get any relevant results from this. R is not a program where you throw numbers in there and wait for magical results to appear, you probably already know this. It's just a tool where we test statistical hypotheses. First of all, we need a question to work on. We then need to thoroughly delineate how we will gather the data. An online survey is just too vague. A good approach that you could do is to pick a set number of players that considers themselves "depressed" and a control group with "mentally stable" players. You could even make it into three classes adding a group of players diagnosed with depression. You replicate it in all leagues so you don't end up with a region bias. You can even go further and rather using "the idea of how a player thinks of themselves" you can use the actual division that the player is currently playing. By doing this I think you can control your experiment much better. Regarding statistical analysis, you should first get a good grasp of the profile of your data so you can run the most appropriate tests. Finally, if you do manage to get any results I don't think the average tftv user will be able to interpret the numbers at all.
But don't let yourself get down if everything fails, a good way to actually learn data science is to get destroyed by failed experiments.

well, my honest opinion
I don't think you'll get any relevant results from this. R is not a program where you throw numbers in there and wait for magical results to appear, you probably already know this. It's just a tool where we test statistical hypotheses. First of all, we need a question to work on. We then need to thoroughly delineate how we will gather the data. An online survey is just too vague. A good approach that you could do is to pick a set number of players that considers themselves "depressed" and a control group with "mentally stable" players. You could even make it into three classes adding a group of players diagnosed with depression. You replicate it in all leagues so you don't end up with a region bias. You can even go further and rather using "the idea of how a player thinks of themselves" you can use the actual division that the player is currently playing. By doing this I think you can control your experiment much better. Regarding statistical analysis, you should first get a good grasp of the profile of your data so you can run the most appropriate tests. Finally, if you do manage to get any results I don't think the average tftv user will be able to interpret the numbers at all.
But don't let yourself get down if everything fails, a good way to actually learn data science is to get destroyed by failed experiments.
15
#15
2 Frags +

Good feedback

Good feedback
16
#16
1 Frags +

and even if you control the experiment the way I suggested you can still come across with problems like low n numbers and excluding variables that are having an effect on the observed phenomena.

and even if you control the experiment the way I suggested you can still come across with problems like low [i]n[/i] numbers and excluding variables that are having an effect on the observed phenomena.
17
#17
2 Frags +

https://i.imgur.com/KtbrRCF.png

where in the world did you get this information from? 2015 nostalgia

[img]https://i.imgur.com/KtbrRCF.png[/img]

where in the world did you get this information from? 2015 nostalgia
Please sign in through STEAM to post a comment.