SlideShare a Scribd company logo
1 of 37
Download to read offline
How do People Type on Mobile Devices?
Observations from a Study with 37,000 Volunteers
Kseniia Palin, Anna Maria Feit, Sunjun Kim, Per Ola Kristensson, Antti Oulasvirta
userinterfaces.aalto.fi/typing37k/
MobileHCI 2019 @ Taipei, Taiwan
2
“If you want to be fast, make use of
both your thumbs and turn on
autocorrection, even though it
might be annoying at times,” said
Feit. “And then just keep using it.”
The researchers from Aalto
University in Finland and Cambridge
University, will present their work at
the International Conference on
Human-Computer Interaction with
Mobile Devices and Services – itself
something of a typing test – in
Taipei, Taiwan, on Wednesday.
https://www.theguardian.com/science/2019/oct/02/ready-text-go-typing-speeds-mobiles-rival-keyboard-users
Typing on mobile device,
known facts:
3
Users control typing speed to compromise between
the accuracy and error.
Banovic et al. 2017, N=20
Typing with one-finger is slower than w/ two thumbs.
Azenkot and Zhai 2012, N=32
Average speed of about 32 WPM; 74% used two thumbs.
Buschek et al. 2018, in-the-wild study, N=30
Texted messages are short: 34 keystrokes per session.
Komninos et al. 2018, in-the-wild study, N=12
Intelligent Text Entry (ITE) methods
4
PredictionAutocorrection Gesture
The keyboard
automatically corrects the
errors in the inputted text.
The keyboard provides a list
of predicted words, and the
user selects one.
An entire word is inputted
at once by drawing a shape
on a keyboard.
Intelligent Text Entry (ITE) methods
5
PredictionAutocorrection Gesture
The keyboard
automatically correct the
error in inputted text
The keyboard provides a list
of predicted words, and the
user selects one.
An entire word is inputted
at once by drawing a shape
on a keyboard.
Open question:
How they are useful in practice?
Method
6
Online typing test
Try it: http://typingtest.aalto.fi/
● Collaboration with Typing Master Inc.
http://typingtest.com
● Period: Sep. 2018 – Jan. 2019
● Transcription task
● 15 random phrases
○ Enron mobile email (memorable set), n=400
○ Gigaword Datasets, n=1125
● Logging
○ Keystroke events
○ Browser meta-data
7
Performance feedback
● Details on typing performance
○ Speed (in Word per Minute)
○ Error (uncorrected)
○ The percentile among the population.
● Visible only after input their demographics:
○ Gender, age, country
○ English language fluency
○ Fingers used for typing
○ etc.
8
Try it: http://typingtest.aalto.fi/
Sample and demographics (after filtering)
9
N=1475N=37,370
Dataset and metrics
10
Over 260,000 started the test.
Over 49,000 completed the test.
We conservatively excluded 25% of participants:
● Users who did not use a mobile device
● Age <5 yo, >61 yo (> 2 SD from the mean)
● Typing speed over 200 WPM
● Uncorrected error >25%
● Long break (>5s) within inputting a sentence
⇒ The final dataset: 37,370 participants.
Words per minute
Uncorrected Error Rate
Keystroke per character
# of backspaces
ITE usage
Keystroke duration
Corrected Error Rate
Interkey interval (IKI)
Recognition of ITE
Per-ITE changes can be detected with a rule set
11
[A quick brpe]t = i
t = i+1 [A quick brpem]
t = i+2 [A quick brown ]
t = i+3 [A quick brown fox ]
t = i+4 [A quick brown fox j]
t = i+5 [A quick brown fox jumps ]
Autocorrection
Gesture
Prediction
Confusion matrix
False Positive = 0.7 %
False Negative = 9.1 %
Selected
results
12
Speed: Words per Minute (WPM)
13
Speed: Words per Minute (WPM)
14
Avg= 36.2
SD=13.2
75%ile: 44
Fastest:
85 WPM!
Speed: Words per Minute (WPM)
15
Higher WPM than previous
Reyal 2015, 31
Buschek 2018, 32
Avg= 36.2
SD=13.2
75%ile: 44
Fastest:
85 WPM!
Error rates (uncorrected)
16
Error rates (uncorrected)
17
Avg= 2.34%
SD=2.08
75%ile: 3.1%
Error rates (uncorrected)
18
● Substitution: 55.6%
● Insertion: 11.1%
● Omission: 33.3%
Avg= 2.34%
SD=2.08
75%ile: 3.1%
Usage vs. Age
19
20s spent the most time for
typing on mobile device.
Age group
Speed vs. Age
20
Age group
Teenagers are the fastest.
→ 39.6 WPM
<10 yo are slowest
→ 24.3 WPM
(* not shown in graph)
Except <10 yo, typing speed
gets slower as age increses.
Speed vs. Language skills
21
Language skill must be
considered when conducting a
text-entry study.
Language experience affect the
typing speed.
(if non-native English users)
Q: How often do you type in English?
Posture
22
Most participants (74%) use both
thumbs to type.
= Buschek et al. 2018
Speed vs. posture
23
Using two fingers is faster than
one-finger typing.
Azenkot 2013 Ours
Two thumbs 50.0 38.0
One thumb 36.3 29.2
One index 33.8 26
Two-thumbs typing is the fastest.
Azenkot and Zhai 2013, Buschek 2018, + ours
Speed vs. ITE
24
A: Autocorrection
P: Prediction
G: Gesture
Speed vs. ITE
25
A: Autocorrection
P: Prediction
G: Gesture
Autocorrection-only users are
faster than all the others
Speed vs. ITE
26
A: Autocorrection
P: Prediction
G: Gesture
Autocorrection-only users are
faster than all the others
Prediction and Gesture
are no faster than no-ITE
Speed vs. ITE
27
A: Autocorrection
P: Prediction
G: Gesture
Some condition is even slower than no-ITE
Autocorrection-only users are
faster than all the others
Prediction and Gesture
are no faster than no-ITE
Typing performances vs. ITE usage
28
A: Autocorrection
P: Prediction
G: Gesture
Pearson correlation values
Typing performances vs. ITE usage
29
A: Autocorrection
P: Prediction
G: Gesture
Pearson correlation values
With more
autocorrections,
the speed gets faster.
Typing performances vs. ITE usage
30
A: Autocorrection
P: Prediction
G: Gesture
Pearson correlation values
With more predictions,
the speed gets slower.
With more
autocorrections,
the speed gets faster.
Typing performances vs. ITE usage
31
A: Autocorrection
P: Prediction
G: Gesture
Pearson correlation values
With more predictions,
the speed gets slower.
With more
autocorrections,
the speed gets faster.
ITEs help slower typists
to have less mistakes.
Typing performances vs. ITE usage
32
A: Autocorrection
P: Prediction
G: Gesture
Pearson correlation values
With more predictions,
the speed gets slower.
With more
autocorrections,
the speed gets faster.
ITEs help slower typists
to have less mistakes.
Prediction and
gesture reduce
keystroke (KSPC)
Conclusion
33
Intelligent Text Entry (ITE) methods contribute
to mobile typing differently.
● Correlations
positive: autocorrection and speed
negative: prediction and speed
● All ITE methods help slow users to
reduce errors.
Typing on mobile device is slow and error prone.
Teenagers have the fastest typing speed.
Two-finger typing is significantly faster than one-finger typing.
Main take-aways
34
Confirmed!
Limitations
35
Sampling bias
● Self-selection bias
● Population bias: western, young, more technology-affined group
● Low proportion of gesture-only users (1.9%)
Imprecision in web-based logging for mobile keystroke events
● Soft keyboard doesn’t transfer touch events to keystroke events as-is.
○ Usually, a set of key-down & key-up events are sent together when touch-up occurs.
● The usage of ITEs were inferred from input text, not directly from the keyboard.
Data
● Raw data (274k participants, 1M sentences, 79M input events).
● Processed data (37k participants, 564k sentences, 27M input events).
Code
● Implementation of the online typing test.
Analysis
● SQL and python scripts used for analyzing and visualizing the data.
● Statistic analysis results.
Public release: The full dataset
36
userinterfaces.aalto.fi/typing37k/
37
How do People Type on Mobile Devices?
Observations from a Study with 37,000 Volunteers
Kseniia Palin, Anna Maria Feit, Sunjun Kim, Per Ola Kristensson, Antti Oulasvirta
New observations
● The first large-scale study with the ITEs.
● Correlations between ITEs and typing speed.
○ Autocorrection: positive.
○ Prediction: negative.
● Novice users get benefits from ITEs for
producing less errors.
Dataset contribution
● 27 million keystrokes from
37k participants.
● Code and analysis scripts
● WPM, error rate, etc.
● All unfiltered raw data from
260k participants.
userinterfaces.aalto.fi/typing37k/

More Related Content

Similar to How do people type on mobile devices? observations from a study with 37,000 volunteers / MobileHCI 2019

Fat Finger - Use of contact Size as a form of simulated pressure.
Fat Finger - Use of contact Size as a form of simulated pressure.Fat Finger - Use of contact Size as a form of simulated pressure.
Fat Finger - Use of contact Size as a form of simulated pressure.Evangelos Tzemis
 
Case Study - TapAssist
Case Study - TapAssistCase Study - TapAssist
Case Study - TapAssistGrace Chien
 
Measuring Efficiency of Use in a Web-based EMR Developed for Malawi: Lessons ...
Measuring Efficiency of Use in a Web-based EMR Developed for Malawi: Lessons ...Measuring Efficiency of Use in a Web-based EMR Developed for Malawi: Lessons ...
Measuring Efficiency of Use in a Web-based EMR Developed for Malawi: Lessons ...Gunther Eysenbach
 
IRJET- Automatic Attendance System using Arduino and GSM Module
IRJET- Automatic Attendance System using Arduino and GSM ModuleIRJET- Automatic Attendance System using Arduino and GSM Module
IRJET- Automatic Attendance System using Arduino and GSM ModuleIRJET Journal
 
IRJET - Mutecom using Tensorflow-Keras Model
IRJET - Mutecom using Tensorflow-Keras ModelIRJET - Mutecom using Tensorflow-Keras Model
IRJET - Mutecom using Tensorflow-Keras ModelIRJET Journal
 
Sign Language Translator
Sign Language TranslatorSign Language Translator
Sign Language TranslatorManjari Raj
 
IRJET- Vision Based Sign Language by using Matlab
IRJET- Vision Based Sign Language by using MatlabIRJET- Vision Based Sign Language by using Matlab
IRJET- Vision Based Sign Language by using MatlabIRJET Journal
 
Live Sign Language Translation: A Survey
Live Sign Language Translation: A SurveyLive Sign Language Translation: A Survey
Live Sign Language Translation: A SurveyIRJET Journal
 
Sign language recognition for deaf and dumb people
Sign language recognition for deaf and dumb peopleSign language recognition for deaf and dumb people
Sign language recognition for deaf and dumb peopleeSAT Publishing House
 
178 - A replicated study on duplicate detection: Using Apache Lucene to searc...
178 - A replicated study on duplicate detection: Using Apache Lucene to searc...178 - A replicated study on duplicate detection: Using Apache Lucene to searc...
178 - A replicated study on duplicate detection: Using Apache Lucene to searc...ESEM 2014
 
INVESTIGATING & IMPROVING THE RELIABILITY AND REPEATABILITY OF KEYSTROKE DYNA...
INVESTIGATING & IMPROVING THE RELIABILITY AND REPEATABILITY OF KEYSTROKE DYNA...INVESTIGATING & IMPROVING THE RELIABILITY AND REPEATABILITY OF KEYSTROKE DYNA...
INVESTIGATING & IMPROVING THE RELIABILITY AND REPEATABILITY OF KEYSTROKE DYNA...IJNSA Journal
 
Electronic Glove: A Teaching AID for the Hearing Impaired
Electronic Glove: A Teaching AID for the Hearing ImpairedElectronic Glove: A Teaching AID for the Hearing Impaired
Electronic Glove: A Teaching AID for the Hearing ImpairedIJECEIAES
 
IRJET- Hand Gesture Recognition for Deaf and Dumb
IRJET- Hand Gesture Recognition for Deaf and DumbIRJET- Hand Gesture Recognition for Deaf and Dumb
IRJET- Hand Gesture Recognition for Deaf and DumbIRJET Journal
 
Objective comparisons of Exploration and Gesture based touchscreen keyboards ...
Objective comparisons of Exploration and Gesture based touchscreen keyboards ...Objective comparisons of Exploration and Gesture based touchscreen keyboards ...
Objective comparisons of Exploration and Gesture based touchscreen keyboards ...Adit Gupta
 
Hand Gesture Recognition using OpenCV and Python
Hand Gesture Recognition using OpenCV and PythonHand Gesture Recognition using OpenCV and Python
Hand Gesture Recognition using OpenCV and Pythonijtsrd
 
iSenseStress: Assessing Stress Through Human-Smartphone Interaction Analysis
iSenseStress: Assessing Stress Through Human-Smartphone Interaction AnalysisiSenseStress: Assessing Stress Through Human-Smartphone Interaction Analysis
iSenseStress: Assessing Stress Through Human-Smartphone Interaction AnalysisKatarzyna Wac & The QoL Lab
 
Smart pen(by ankit)
Smart pen(by ankit)Smart pen(by ankit)
Smart pen(by ankit)rankitreddy
 
Isce osk tcs
Isce osk tcsIsce osk tcs
Isce osk tcsArpan Pal
 

Similar to How do people type on mobile devices? observations from a study with 37,000 volunteers / MobileHCI 2019 (20)

Fat Finger - Use of contact Size as a form of simulated pressure.
Fat Finger - Use of contact Size as a form of simulated pressure.Fat Finger - Use of contact Size as a form of simulated pressure.
Fat Finger - Use of contact Size as a form of simulated pressure.
 
Case Study - TapAssist
Case Study - TapAssistCase Study - TapAssist
Case Study - TapAssist
 
Measuring Efficiency of Use in a Web-based EMR Developed for Malawi: Lessons ...
Measuring Efficiency of Use in a Web-based EMR Developed for Malawi: Lessons ...Measuring Efficiency of Use in a Web-based EMR Developed for Malawi: Lessons ...
Measuring Efficiency of Use in a Web-based EMR Developed for Malawi: Lessons ...
 
IRJET- Automatic Attendance System using Arduino and GSM Module
IRJET- Automatic Attendance System using Arduino and GSM ModuleIRJET- Automatic Attendance System using Arduino and GSM Module
IRJET- Automatic Attendance System using Arduino and GSM Module
 
IRJET - Mutecom using Tensorflow-Keras Model
IRJET - Mutecom using Tensorflow-Keras ModelIRJET - Mutecom using Tensorflow-Keras Model
IRJET - Mutecom using Tensorflow-Keras Model
 
AI Screen Presence
AI Screen PresenceAI Screen Presence
AI Screen Presence
 
Sign Language Translator
Sign Language TranslatorSign Language Translator
Sign Language Translator
 
IRJET- Vision Based Sign Language by using Matlab
IRJET- Vision Based Sign Language by using MatlabIRJET- Vision Based Sign Language by using Matlab
IRJET- Vision Based Sign Language by using Matlab
 
Live Sign Language Translation: A Survey
Live Sign Language Translation: A SurveyLive Sign Language Translation: A Survey
Live Sign Language Translation: A Survey
 
Sign language recognition for deaf and dumb people
Sign language recognition for deaf and dumb peopleSign language recognition for deaf and dumb people
Sign language recognition for deaf and dumb people
 
178 - A replicated study on duplicate detection: Using Apache Lucene to searc...
178 - A replicated study on duplicate detection: Using Apache Lucene to searc...178 - A replicated study on duplicate detection: Using Apache Lucene to searc...
178 - A replicated study on duplicate detection: Using Apache Lucene to searc...
 
INVESTIGATING & IMPROVING THE RELIABILITY AND REPEATABILITY OF KEYSTROKE DYNA...
INVESTIGATING & IMPROVING THE RELIABILITY AND REPEATABILITY OF KEYSTROKE DYNA...INVESTIGATING & IMPROVING THE RELIABILITY AND REPEATABILITY OF KEYSTROKE DYNA...
INVESTIGATING & IMPROVING THE RELIABILITY AND REPEATABILITY OF KEYSTROKE DYNA...
 
Electronic Glove: A Teaching AID for the Hearing Impaired
Electronic Glove: A Teaching AID for the Hearing ImpairedElectronic Glove: A Teaching AID for the Hearing Impaired
Electronic Glove: A Teaching AID for the Hearing Impaired
 
IRJET- Hand Gesture Recognition for Deaf and Dumb
IRJET- Hand Gesture Recognition for Deaf and DumbIRJET- Hand Gesture Recognition for Deaf and Dumb
IRJET- Hand Gesture Recognition for Deaf and Dumb
 
Objective comparisons of Exploration and Gesture based touchscreen keyboards ...
Objective comparisons of Exploration and Gesture based touchscreen keyboards ...Objective comparisons of Exploration and Gesture based touchscreen keyboards ...
Objective comparisons of Exploration and Gesture based touchscreen keyboards ...
 
Gyropen ppt
Gyropen pptGyropen ppt
Gyropen ppt
 
Hand Gesture Recognition using OpenCV and Python
Hand Gesture Recognition using OpenCV and PythonHand Gesture Recognition using OpenCV and Python
Hand Gesture Recognition using OpenCV and Python
 
iSenseStress: Assessing Stress Through Human-Smartphone Interaction Analysis
iSenseStress: Assessing Stress Through Human-Smartphone Interaction AnalysisiSenseStress: Assessing Stress Through Human-Smartphone Interaction Analysis
iSenseStress: Assessing Stress Through Human-Smartphone Interaction Analysis
 
Smart pen(by ankit)
Smart pen(by ankit)Smart pen(by ankit)
Smart pen(by ankit)
 
Isce osk tcs
Isce osk tcsIsce osk tcs
Isce osk tcs
 

Recently uploaded

Developing An App To Navigate The Roads of Brazil
Developing An App To Navigate The Roads of BrazilDeveloping An App To Navigate The Roads of Brazil
Developing An App To Navigate The Roads of BrazilV3cube
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsMaria Levchenko
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Allon Mureinik
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerThousandEyes
 
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfThe Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfEnterprise Knowledge
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountPuma Security, LLC
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationSafe Software
 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024The Digital Insurer
 
How to convert PDF to text with Nanonets
How to convert PDF to text with NanonetsHow to convert PDF to text with Nanonets
How to convert PDF to text with Nanonetsnaman860154
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking MenDelhi Call girls
 
Kalyanpur ) Call Girls in Lucknow Finest Escorts Service 🍸 8923113531 🎰 Avail...
Kalyanpur ) Call Girls in Lucknow Finest Escorts Service 🍸 8923113531 🎰 Avail...Kalyanpur ) Call Girls in Lucknow Finest Escorts Service 🍸 8923113531 🎰 Avail...
Kalyanpur ) Call Girls in Lucknow Finest Escorts Service 🍸 8923113531 🎰 Avail...gurkirankumar98700
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdfhans926745
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxMalak Abu Hammad
 
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...Igalia
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreternaman860154
 
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of ServiceCNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Servicegiselly40
 
Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024The Digital Insurer
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)Gabriella Davis
 
WhatsApp 9892124323 ✓Call Girls In Kalyan ( Mumbai ) secure service
WhatsApp 9892124323 ✓Call Girls In Kalyan ( Mumbai ) secure serviceWhatsApp 9892124323 ✓Call Girls In Kalyan ( Mumbai ) secure service
WhatsApp 9892124323 ✓Call Girls In Kalyan ( Mumbai ) secure servicePooja Nehwal
 
Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityBoost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityPrincipled Technologies
 

Recently uploaded (20)

Developing An App To Navigate The Roads of Brazil
Developing An App To Navigate The Roads of BrazilDeveloping An App To Navigate The Roads of Brazil
Developing An App To Navigate The Roads of Brazil
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed texts
 
Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)Injustice - Developers Among Us (SciFiDevCon 2024)
Injustice - Developers Among Us (SciFiDevCon 2024)
 
How to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected WorkerHow to Troubleshoot Apps for the Modern Connected Worker
How to Troubleshoot Apps for the Modern Connected Worker
 
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdfThe Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
The Role of Taxonomy and Ontology in Semantic Layers - Heather Hedden.pdf
 
Breaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path MountBreaking the Kubernetes Kill Chain: Host Path Mount
Breaking the Kubernetes Kill Chain: Host Path Mount
 
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time AutomationFrom Event to Action: Accelerate Your Decision Making with Real-Time Automation
From Event to Action: Accelerate Your Decision Making with Real-Time Automation
 
Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024Axa Assurance Maroc - Insurer Innovation Award 2024
Axa Assurance Maroc - Insurer Innovation Award 2024
 
How to convert PDF to text with Nanonets
How to convert PDF to text with NanonetsHow to convert PDF to text with Nanonets
How to convert PDF to text with Nanonets
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
 
Kalyanpur ) Call Girls in Lucknow Finest Escorts Service 🍸 8923113531 🎰 Avail...
Kalyanpur ) Call Girls in Lucknow Finest Escorts Service 🍸 8923113531 🎰 Avail...Kalyanpur ) Call Girls in Lucknow Finest Escorts Service 🍸 8923113531 🎰 Avail...
Kalyanpur ) Call Girls in Lucknow Finest Escorts Service 🍸 8923113531 🎰 Avail...
 
[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf[2024]Digital Global Overview Report 2024 Meltwater.pdf
[2024]Digital Global Overview Report 2024 Meltwater.pdf
 
The Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptxThe Codex of Business Writing Software for Real-World Solutions 2.pptx
The Codex of Business Writing Software for Real-World Solutions 2.pptx
 
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
 
Presentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreterPresentation on how to chat with PDF using ChatGPT code interpreter
Presentation on how to chat with PDF using ChatGPT code interpreter
 
CNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of ServiceCNv6 Instructor Chapter 6 Quality of Service
CNv6 Instructor Chapter 6 Quality of Service
 
Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024
 
A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)A Domino Admins Adventures (Engage 2024)
A Domino Admins Adventures (Engage 2024)
 
WhatsApp 9892124323 ✓Call Girls In Kalyan ( Mumbai ) secure service
WhatsApp 9892124323 ✓Call Girls In Kalyan ( Mumbai ) secure serviceWhatsApp 9892124323 ✓Call Girls In Kalyan ( Mumbai ) secure service
WhatsApp 9892124323 ✓Call Girls In Kalyan ( Mumbai ) secure service
 
Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityBoost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivity
 

How do people type on mobile devices? observations from a study with 37,000 volunteers / MobileHCI 2019

  • 1. How do People Type on Mobile Devices? Observations from a Study with 37,000 Volunteers Kseniia Palin, Anna Maria Feit, Sunjun Kim, Per Ola Kristensson, Antti Oulasvirta userinterfaces.aalto.fi/typing37k/ MobileHCI 2019 @ Taipei, Taiwan
  • 2. 2 “If you want to be fast, make use of both your thumbs and turn on autocorrection, even though it might be annoying at times,” said Feit. “And then just keep using it.” The researchers from Aalto University in Finland and Cambridge University, will present their work at the International Conference on Human-Computer Interaction with Mobile Devices and Services – itself something of a typing test – in Taipei, Taiwan, on Wednesday. https://www.theguardian.com/science/2019/oct/02/ready-text-go-typing-speeds-mobiles-rival-keyboard-users
  • 3. Typing on mobile device, known facts: 3 Users control typing speed to compromise between the accuracy and error. Banovic et al. 2017, N=20 Typing with one-finger is slower than w/ two thumbs. Azenkot and Zhai 2012, N=32 Average speed of about 32 WPM; 74% used two thumbs. Buschek et al. 2018, in-the-wild study, N=30 Texted messages are short: 34 keystrokes per session. Komninos et al. 2018, in-the-wild study, N=12
  • 4. Intelligent Text Entry (ITE) methods 4 PredictionAutocorrection Gesture The keyboard automatically corrects the errors in the inputted text. The keyboard provides a list of predicted words, and the user selects one. An entire word is inputted at once by drawing a shape on a keyboard.
  • 5. Intelligent Text Entry (ITE) methods 5 PredictionAutocorrection Gesture The keyboard automatically correct the error in inputted text The keyboard provides a list of predicted words, and the user selects one. An entire word is inputted at once by drawing a shape on a keyboard. Open question: How they are useful in practice?
  • 7. Online typing test Try it: http://typingtest.aalto.fi/ ● Collaboration with Typing Master Inc. http://typingtest.com ● Period: Sep. 2018 – Jan. 2019 ● Transcription task ● 15 random phrases ○ Enron mobile email (memorable set), n=400 ○ Gigaword Datasets, n=1125 ● Logging ○ Keystroke events ○ Browser meta-data 7
  • 8. Performance feedback ● Details on typing performance ○ Speed (in Word per Minute) ○ Error (uncorrected) ○ The percentile among the population. ● Visible only after input their demographics: ○ Gender, age, country ○ English language fluency ○ Fingers used for typing ○ etc. 8 Try it: http://typingtest.aalto.fi/
  • 9. Sample and demographics (after filtering) 9 N=1475N=37,370
  • 10. Dataset and metrics 10 Over 260,000 started the test. Over 49,000 completed the test. We conservatively excluded 25% of participants: ● Users who did not use a mobile device ● Age <5 yo, >61 yo (> 2 SD from the mean) ● Typing speed over 200 WPM ● Uncorrected error >25% ● Long break (>5s) within inputting a sentence ⇒ The final dataset: 37,370 participants. Words per minute Uncorrected Error Rate Keystroke per character # of backspaces ITE usage Keystroke duration Corrected Error Rate Interkey interval (IKI)
  • 11. Recognition of ITE Per-ITE changes can be detected with a rule set 11 [A quick brpe]t = i t = i+1 [A quick brpem] t = i+2 [A quick brown ] t = i+3 [A quick brown fox ] t = i+4 [A quick brown fox j] t = i+5 [A quick brown fox jumps ] Autocorrection Gesture Prediction Confusion matrix False Positive = 0.7 % False Negative = 9.1 %
  • 13. Speed: Words per Minute (WPM) 13
  • 14. Speed: Words per Minute (WPM) 14 Avg= 36.2 SD=13.2 75%ile: 44 Fastest: 85 WPM!
  • 15. Speed: Words per Minute (WPM) 15 Higher WPM than previous Reyal 2015, 31 Buschek 2018, 32 Avg= 36.2 SD=13.2 75%ile: 44 Fastest: 85 WPM!
  • 17. Error rates (uncorrected) 17 Avg= 2.34% SD=2.08 75%ile: 3.1%
  • 18. Error rates (uncorrected) 18 ● Substitution: 55.6% ● Insertion: 11.1% ● Omission: 33.3% Avg= 2.34% SD=2.08 75%ile: 3.1%
  • 19. Usage vs. Age 19 20s spent the most time for typing on mobile device. Age group
  • 20. Speed vs. Age 20 Age group Teenagers are the fastest. → 39.6 WPM <10 yo are slowest → 24.3 WPM (* not shown in graph) Except <10 yo, typing speed gets slower as age increses.
  • 21. Speed vs. Language skills 21 Language skill must be considered when conducting a text-entry study. Language experience affect the typing speed. (if non-native English users) Q: How often do you type in English?
  • 22. Posture 22 Most participants (74%) use both thumbs to type. = Buschek et al. 2018
  • 23. Speed vs. posture 23 Using two fingers is faster than one-finger typing. Azenkot 2013 Ours Two thumbs 50.0 38.0 One thumb 36.3 29.2 One index 33.8 26 Two-thumbs typing is the fastest. Azenkot and Zhai 2013, Buschek 2018, + ours
  • 24. Speed vs. ITE 24 A: Autocorrection P: Prediction G: Gesture
  • 25. Speed vs. ITE 25 A: Autocorrection P: Prediction G: Gesture Autocorrection-only users are faster than all the others
  • 26. Speed vs. ITE 26 A: Autocorrection P: Prediction G: Gesture Autocorrection-only users are faster than all the others Prediction and Gesture are no faster than no-ITE
  • 27. Speed vs. ITE 27 A: Autocorrection P: Prediction G: Gesture Some condition is even slower than no-ITE Autocorrection-only users are faster than all the others Prediction and Gesture are no faster than no-ITE
  • 28. Typing performances vs. ITE usage 28 A: Autocorrection P: Prediction G: Gesture Pearson correlation values
  • 29. Typing performances vs. ITE usage 29 A: Autocorrection P: Prediction G: Gesture Pearson correlation values With more autocorrections, the speed gets faster.
  • 30. Typing performances vs. ITE usage 30 A: Autocorrection P: Prediction G: Gesture Pearson correlation values With more predictions, the speed gets slower. With more autocorrections, the speed gets faster.
  • 31. Typing performances vs. ITE usage 31 A: Autocorrection P: Prediction G: Gesture Pearson correlation values With more predictions, the speed gets slower. With more autocorrections, the speed gets faster. ITEs help slower typists to have less mistakes.
  • 32. Typing performances vs. ITE usage 32 A: Autocorrection P: Prediction G: Gesture Pearson correlation values With more predictions, the speed gets slower. With more autocorrections, the speed gets faster. ITEs help slower typists to have less mistakes. Prediction and gesture reduce keystroke (KSPC)
  • 34. Intelligent Text Entry (ITE) methods contribute to mobile typing differently. ● Correlations positive: autocorrection and speed negative: prediction and speed ● All ITE methods help slow users to reduce errors. Typing on mobile device is slow and error prone. Teenagers have the fastest typing speed. Two-finger typing is significantly faster than one-finger typing. Main take-aways 34 Confirmed!
  • 35. Limitations 35 Sampling bias ● Self-selection bias ● Population bias: western, young, more technology-affined group ● Low proportion of gesture-only users (1.9%) Imprecision in web-based logging for mobile keystroke events ● Soft keyboard doesn’t transfer touch events to keystroke events as-is. ○ Usually, a set of key-down & key-up events are sent together when touch-up occurs. ● The usage of ITEs were inferred from input text, not directly from the keyboard.
  • 36. Data ● Raw data (274k participants, 1M sentences, 79M input events). ● Processed data (37k participants, 564k sentences, 27M input events). Code ● Implementation of the online typing test. Analysis ● SQL and python scripts used for analyzing and visualizing the data. ● Statistic analysis results. Public release: The full dataset 36 userinterfaces.aalto.fi/typing37k/
  • 37. 37 How do People Type on Mobile Devices? Observations from a Study with 37,000 Volunteers Kseniia Palin, Anna Maria Feit, Sunjun Kim, Per Ola Kristensson, Antti Oulasvirta New observations ● The first large-scale study with the ITEs. ● Correlations between ITEs and typing speed. ○ Autocorrection: positive. ○ Prediction: negative. ● Novice users get benefits from ITEs for producing less errors. Dataset contribution ● 27 million keystrokes from 37k participants. ● Code and analysis scripts ● WPM, error rate, etc. ● All unfiltered raw data from 260k participants. userinterfaces.aalto.fi/typing37k/