Minecraft PC IP: play.cubecraft.net

Fesa

Quality Assurance
Team CubeCraft
🧪 Quality Assurance
Oct 17, 2016
1,505
4,014
349
Yoitsu
Pronouns
She/Her
Hi!

If you’re somewhat active on the CubeCraft discord you’ll most likely now I’ve spend part of last month making graphs about activity in the discord. This thread will be dedicated towards Staff Help since that has been running for a month minus one day, and the other haven’t. I’ll likely make a second thread when the rest has been running for a month since I like doing it :D
I've littered in some questions during my own observations, these are written in italic!

Before you make conclusions with the graphs, please read the remarks and notes very well. There are subtle difference between what you might think a graph means and what it actually means. I'll put all remarks and notes in quotes like so:
This is a very important remark, don't forget me!​

Bit of explanation on the graphics site;
Every bar graph has colour, these indicate in which 20% region of the date the bar is.
  • 80% to 100% is red
  • 60% to 80% is orange
  • 40% to 60% is yellow
  • 20% to 40% is green
  • 0% to 20% is blue
Graphs with continuous lines have one colour, with no meaning.

First remarks
Every section will have two graphs. One with date from 2022-05-11 and one from 2022-05-28; the promotion date of the current helpers.​
Some names have empty rectangle characters, this means that matplotlib doesn’t know the character. You’ll have to guess, or ask, which staff member this is.​

Quick overview of the last month​

Full month data
Figure_2.png

It is very obvious that staff helped is mainly staffed by the same people. And that Warz is an absolute mad-lad with more than a thousand replies in staff help over the last month, with second up runner only having 57,3% of Warz' total replies. Using the Percentages tag we see that Warz has 7,0% off the total messages. Cathie comes close with 6,5% since she has received Helper rank. Which staff member would you love to make an appearance in the top 40% next month? Don't tag them!

Who is asking questions and when are questions being asked?​

Figure_3.png
Figure_4.png

The role attribute was implemented only later, so the graphs are made up of the same data.​
Role Distribution also includes members of the same team. Keep in mind that they all have their respective forums and version roles.​
The Hourly Activity graphs included non staff member data only.​
Shifting the "Forums" role percentage to not include members of the staff team makes players with no roles the most frequent seekers for help in Staff Help with a grand total of 44,4% of all messages. Players with a connected Java account are slight more in need of answer in contrast to players with a connected Bedrock account with 31,3% to 30,9% of all messages. Both the highest rank for each version sends to most messages in staff help (these are Obsidian rank and Vip lvl 5 for Java and Bedrock respectively). My biggest surprise was the Admin team having more messages than the bottom three ranks on Java edition! What surprised you the most?

Looking at the longest time frame, the most active time lays between 21h00 and 22h00. And overal are most questions asked after four in the afternoon until around one in the morning. This might be explained by EU players coming home from school/work/... and NA players waking up. But do you have an explanation why six to seven in the morning spiked the last two weeks?


Staff Help is so slow! But how slow is it really?​


Figure_5.png
Figure_6.png

No graphs from the last two weeks are included because of the lack of data.​
Extremely high spikes are contributed to a very low amount of questions during that time (see last sections). These should not be taken into account!​
Two messages are ignored in order to compensate for "Thank you" and confirmation messages from members.​
The Hourly Activity graph includes staff member data only.​
It is quite obvious that the graphs seems to be inverses of each other. Which is to be expected, less staff members and questions results in longer wait times. During the most active hours, as concluded in the last section, the average wait time would be 9min8s! This is more than acceptable in my eyes, so players during these time are not allowed to complain hooray!

But oh no! The downtime is horrible, while the amount of messages from seven to eight in the morning is only five times less than from four to five in the evening. The reply times has been multiplied by 8,5!!
We can safely conclude that EU staff members rule! Do you have any other insight?
Quize time: What would be the worse time to ask a question down to 10min specific (use the left graph). Ignoring the 3 highest spikes as per the remark.


No more graphs 😭

This has been amazingly fun to make, so I'll be doing it again next month with an all time (I.e. from 2022-05-11) and a last month recap. Including and English General and all other channels section!
Do you have some extra idea's for new graphs? Please do tell!
Any thing you noticed and I didn't? Tell us!

I still want more graphs!​

I can also make graphs like the one below for a specific staff member, if you are interested!
Figure_7.png


How does it work?​

Read the code on my GitHub!


Brought with love by Fesa and the Python programming language ♥️
 
Last edited:

Stijnnn_

Well-Known Member
Jan 4, 2021
404
567
129
Bamboo forest
Hi!

If you’re somewhat active on the CubeCraft discord you’ll most likely now I’ve spend part of last month making graphs about activity in the discord. This thread will be dedicated towards Staff Help since that has been running for a month minus one day, and the other haven’t. I’ll likely make a second thread when the rest has been running for a month since I like doing it :D
I've littered in some questions during my own observations, these are written in italic!

Before you make conclusions with the graphs, please read the remarks and notes very well. There are subtle difference between what you might think a graph means and what it actually means. I'll put all remarks and notes in quotes like so:


Bit of explanation on the graphics site;
Every bar graph has colour, these indicate in which 20% region of the date the bar is.
  • 80% to 100% is red
  • 60% to 80% is orange
  • 40% to 60% is yellow
  • 20% to 40% is green
  • 0% to 20% is blue
Graphs with continuous lines have one colour, with no meaning.

First remarks



Quick overview of the last month​

View attachment 211650 View attachment 211654
It is very obvious that staff helped is mainly staffed by the same people. And that Warz is an absolute mad-lad with more than a thousand replies in staff help over the last month, with second up runner only having 57,3% of Warz' total replies. Using the Percentages tag we see that Warz has 7,0% off the total messages. Cathie comes close with 6,5% since she has received Helper rank. Which staff member would you love to make an appearance in the top 40% next month? Don't tag them!

Who is asking questions and when are questions being asked?​

View attachment 211655 View attachment 211656



Shifting the "Forums" role percentage to not include members of the staff team makes players with no roles the most frequent seekers for help in Staff Help with a grand total of 44,4% of all messages. Players with a connected Java account are slight more in need of answer in contrast to players with a connected Bedrock account with 31,3% to 30,9% of all messages. Both the highest rank for each version sends to most messages in staff help (these are Obsidian rank and Vip lvl 5 for Java and Bedrock respectively). My biggest surprise was the Admin team having more messages than the bottom three ranks on Java edition! What surprised you the most?

Looking at the longest time frame, the most active time lays between 21h00 and 22h00. And overal are most questions asked after four in the afternoon until around one in the morning. This might be explained by EU players coming home from school/work/... and NA players waking up. But do you have an explanation why six to seven in the morning spiked the last two weeks?


Staff Help is so slow! But how slow is it really?​


View attachment 211657 View attachment 211658




It is quite obvious that the graphs seems to be inverses of each other. Which is to be expected, less staff members and questions results in longer wait times. During the most active hours, as concluded in the last section, the average wait time would be 9min8s! This is more than acceptable in my eyes, so players during these time are not allowed to complain hooray!

But oh no! The downtime is horrible, while the amount of messages from seven to eight in the morning is only five times less than from four to five in the evening. The reply times has been multiplied by 8,5!!
We can safely conclude that EU staff members rule! Do you have any other insight?
Quize time: What would be the worse time to ask a question down to 10min specific (use the left graph). Ignoring the 3 highest spikes as per the remark.


No more graphs 😭

This has been amazingly fun to make, so I'll be doing it again next month with an all time (I.e. from 2022-05-11) and a last month recap. Including and English General and all other channels section!
Do you have some extra idea's for new graphs? Please do tell!
Any thing you noticed and I didn't? Tell us!

I still want more graphs!​

I can also make graphs like the one below for a specific staff member, if you are interested!
View attachment 211659

How does it work?​

Read the code on my GitHub!


Brought with love by Fesa and the Python programming language ♥️
this is so cool! you should totally do this more often if you have the time/motivation for it
 
  • Like
Reactions: Fesa

tiger101tiger

Well-Known Member
Jul 9, 2014
798
371
138
USA

How does it work?​

Read the code on my GitHub!
Pretty cool project. Can I have a link to your GitHub? I'm mainly interested in how you determined if a question was answered, and which package you used to make those pretty graphics.
 

Fesa

Quality Assurance
Team CubeCraft
🧪 Quality Assurance
Oct 17, 2016
1,505
4,014
349
Yoitsu
Pronouns
She/Her
Pretty cool project. Can I have a link to your GitHub? I'm mainly interested in how you determined if a question was answered, and which package you used to make those pretty graphics.

There is no way for me to have acces to message content, so it just takes the time until a staff reply. I believe on average, this would be a quite alright approximation for reply time.

The module used is matplotlib.
 

SanCookie

Gameplay Engineer
Team CubeCraft
🖌️ Sr. Designer
Mar 18, 2017
2,304
12,680
498
22
the selfie area
bit.ly
Pronouns
He/Him
Now that I think about it, you should post the statistics for the rarity of Cubelets, and how long it'd take in theory to get all loot from cubelets, etc.
 
  • Love
Reactions: Fesa

Fesa

Quality Assurance
Team CubeCraft
🧪 Quality Assurance
Oct 17, 2016
1,505
4,014
349
Yoitsu
Pronouns
She/Her
Now that I think about it, you should post the statistics for the rarity of Cubelets, and how long it'd take in theory to get all loot from cubelets, etc.
Would be a most interesting study.
I would have to assume that there is no actual chance for a duplicate and that it is just a uniform choice (per rarity) on an item. Making duplicates more common as you have more items.
Ooh, this could be quite hard to get anything useful from it as I have no idea how many items I had at which point
Good thing I have my own loot database so I can check if there is a different chance per rarity or not even. AAAA SO COOL IDEA

I might begin trying this evening or after exams :D
 

SanCookie

Gameplay Engineer
Team CubeCraft
🖌️ Sr. Designer
Mar 18, 2017
2,304
12,680
498
22
the selfie area
bit.ly
Pronouns
He/Him
Would be a most interesting study.
I would have to assume that there is no actual chance for a duplicate and that it is just a uniform choice (per rarity) on an item. Making duplicates more common as you have more items.
Ooh, this could be quite hard to get anything useful from it as I have no idea how many items I had at which point
Good thing I have my own loot database so I can check if there is a different chance per rarity or not even. AAAA SO COOL IDEA

I might begin trying this evening or after exams :D
Yeah! It's most likely a sigmoid curve, you could also determine the time it takes to get 50% of all things
I think I sent u data from votelets a while back, can use that as a fun little side thing
 
  • Love
Reactions: Fesa
Members Online

Team online

Latest profile posts

Sergiio12 wrote on LoveAndMafia's profile.
Alejaaaandro
Blom wrote on Lyriie's profile.
Lyriie, sorry for bullying you all the time :o
Ur an amazing staff member <3
mojang nerf mobile players pls
Someone nerfed me because I can't find those hidden chests as easily anymore.. I won't stop trying tho >:c
Top Bottom