Anomaly detection and data imputation within time series
gameF2PUX.pptx
1. Game UX
Shipped games send the signal in the form of data,
We use those data to fine-tune players’ desired paths.
- Karan
2. Desire paths simply take one from point A to point B
(Design flow). These are alternate routes taken by
players instead of the built path because they are more
convenient. Over time, a new “desire path” is created
different from the one actually designed.
Desire paths
3. Data to explain a particular event or phenomenon and
represent the outcome in numerical form. In UX It helps the
Financial success of a solution.
• Find out about player experience at scale? (Outcomes)
• We use that information to influence our design
decisions
• We know when we’ve got it right? Pervasive
justification for design choices.
The Design which is focused on Data called as Data-
Driven Design or Quantitative Design
Why design with Data?
4. Type of Research?
There are two major type of Research
1. Qualitative
Qualitative research relies on data obtained by the
researcher from first-hand observation, interviews,
questionnaires, focus groups, participant-observation,
recordings made in natural settings, documents, case
studies, and artefacts. More:
https://www.slideshare.net/KamalKanth5/ux-research-
174838729
2. Quantitative
All about numbers
5. We use Playtestcloud, In one way it’s very helpful to
find about players’ desired path but it has it own
cons.
1. Recruit the wrong test participants (Gamers).
2. We can’t iterate on our test script (Run time).
3. We miss Failing participant.
Pitfalls in recruiting for
Qualitative
7. 1. Population, Collection of all data (N -
Parameters).
• If the data is small Eg: MyWhoosh event
data
• More accurate. (Country Population)
• Expensive and difficult to calculate
2. Sample, Sub set of population (n - Statistics)
• Less time and Cheaper
Data type
9. 1. Random
2. Representative
• Stratified sampling (Dividing the target
population Eg: Payer, non-payer)
• Volunteer sampling. (Player who gave 5star
rating)
• Opportunity sampling. (Player who is opt-in
for particular time)
Statistics
10. The categorical data depicts the success and failure
rate. It’s important to know whether a difference is
large enough to be meaningful or simply random
chance.
The p-value helps us to decide if we have strong
enough evidence to make conclusions about our
data.
Playsimple UX Method:
1. Set a Hypothesis
2. Define Individuals
3. Set expected values
Make sense of data -
Statistical Significance
11. The collection and examination of numerical data to define, explain, and predict variables.
12. • Basic functionality are FREE, Advance
functionality at a fee.
• Potential to massive scale
• Monitization feature used for most
engaged users
FREEMIUM
13. • Highly engaged users called “Whales” (~
5%)
• Broader the appeal (Ads)
• High level of engagement (need tracking)
• Track product performance
• More likely to monetise the user, When,
why and How
• Continue development based on
ARPDAU
Components