No clusters to visualize
To visualize clusters:
choose data, dimensions you want to cluster and algorithm
To display information about
the cluster please move your
cursor to one of the clusters
202 attempts — 74% of all attempts
Validation
Details of the cluster
100%—
98%—
97%—
95%—
BATTER to MINER
MINER to BATTER
CLIMBER to MINER
BLOCKER to CLIMBER
7
195
9% of attempts
might belong to
a different cluster
List of attempts
Cluster 1
Save Group
40 attempts — 15% of all attempts
Validation
Details of the cluster
100%—
100%—
95%—
92%—
CLIMBER to CLIMBER
JUMPER to JUMPER
PLATFORMER to PLATFORM..
MINER to PLATFORMER
3
37
7,5% of attempts
might belong to
a different cluster
List of attempts
Cluster 2
Save Group
Cluster 3
30 attempts — 11% of all attempts
Validation
Details of the cluster
95%—
92%—
87%—
85%—
MINER to PLATFORMER
PLATFORMER to MINER
MINER to MINER
JUMPER to PLATFORMER
1
29
3,4% of attempts
might belong to
a different cluster
List of attempts
Save Group
K-means clustering of Transitions between 2 skills
272 attempts, 45 dimensions — Grouped by cluster
Cluster 1
Cluster 2
Cluster 3
To display information about the group please move your cursor to one of the groups
24 attempts — 12% of all attempts
Details of the group
30%
Cluster 2
80%
of total
Cluster 1
37%
57%
Cluster 3
6%
2%
of group
List of attempts
Output: 1 [won]
250 attempts — 88% of all attempts
Details of the group
70%
Cluster 2
20%
of total
Cluster 1
24%
59%
Cluster 3
17%
98%
of group
List of attempts
Output: 1 [lost]
K-means clustering of Transitions between 2 skills
272 attempts, 45 dimensions — Grouped by outcome
Cluster 1
Cluster 2
Cluster 3
Output: 0 [lost]
Output: 1 [won]
K-means clustering of Transitions between 2 skills
272 attempts, 45 dimensions — Grouped by cluster
8 attempts
Details of the Player 2
Cluster 2
Cluster 1
4 — 50%
3 — 37,5%
Cluster 3
1 — 12,5%
List of attempts by cluster
Player 2 is highlighted
Cluster 3
Cluster 2
Cluster 1
1
2
3
4
5
6
7
8
Attempt
Cluster 1
Cluster 2
Cluster 3
ALGORITHM SETTINGS
DATA
15
272
Players [1, 2, 3,...]
Attempts [1, 2, 3,...]
Skills [CLIMBER, MINER, ...]
2552
2552
Timestamps [11/07/16 10:16:08, ...]
2
Outputs [0, 1]
Edit Data
Upload New Data
5
Extra options [skills_id, which_lix, ...]
You selected 43 dimensions. Choose clustering algorithm:
* — recommended for you dataset
K-MEANS SETTINGS
Number of clusters (k):
3 is recommended for you dataset
Some help!
Reset Settings
Visualize Results
VISUZALIZATION
SETTINGS
No visualization
Group visualization by:
Cluster 1
Cluster 2
Cluster 3
Outcome: 0
Outcome: 1
Highlight group:
Time cluster
Player 2
+ Create a group
FAVOURITES
Cluster times
Wining patterns
+ Save current settings
Heading 2
S U P E R A W E S O M E C L U S T E R I N G T O O L
CLUSTERING SETTINGS
1. Cluster all attempts:
2. Include:
3. Choose dimensions:
Include attempts with
min skills:
max skills:
Settings
+
Total time spent
1 OR 2 DIMENSIONS
Explanation
Cluster attempts by time they took. It is possible to set two dimensions: median times for attempts with outcome 0 and 1 [win and lose]
+
Time
between 2 skills
6 DIMENSIONS
Settings
Include attempts with
min transitions:
max transitions:
Explanation
Cluster times between. This may show you slow-fast attempts.
Each period will be assigned to one of the time-bins: 5s, 10s, 30s, 50s or more.
+
Explanation
You can set more
dimension for the specific parameters
of your data.
For example:
which_lix, lix_saved
Configure specific
dimensions based
on your dataset
+
Number
of each skill
8 DIMENSIONS
Settings
Include attempts with
min skills:
max skills:
Explanation
Cluster attempts based on how often each skill is used.
Each dimension is a skill: CLIMBER, JUMPER, etc.
+
Transitions
between 2 skills
43 DIMENSIONS
Settings
Include attempts with
min transitions:
max transitions:
Explanation
Cluster attempts based on transitions between every two skills.
For example: CLIMBER to JUMPER,
MINER to BATTER,...
+
Common
skills patterns
SET N OF DIMENSIONS
Settings
Include attempts with
min skills:
max skills:
Explanation
Cluster attempts by pattern(s) —sequences of skills — they share.
This clustering may help to find groups with the same pattern of behavior.
Some help!
Some help!
Some help!