The document presents a stage-based model of personal informatics systems composed of 5 stages: Preparation, Collection, Integration, Reflection, and Action. It identifies common barriers that users experience in each stage, such as not having the right tools for collection or not having enough time for reflection. The stages are iterative and problems can cascade between stages. Development of personal informatics systems should consider the whole user experience and balance automation with user control.
1. A Stage-Based Model Ian Li
Anind Dey
of Personal Informatics Systems Jodi Forlizzi
Introduction
People strive to obtain self-knowledge. A class of systems called personal We conducted surveys and interviews with people who collect and reflect
informatics is appearing that help people collect and reflect on personal in- on personal information. We derived a stage-based model of personal in-
formation. However, there is no comprehensive list of problems that users formatics systems composed of five stages (Preparation, Collection, Inte-
experience using these systems, and no guidance for making these sys- gration, Reflection, and Action) and identified barriers in each of the stages.
tems more effective.
http://personalinformatics.org/lab/model
Stages
PREPARATION COLLECTION INTEGRATION REFLECTION ACTION
The Preparation Stage concerns The Collection Stage is the time when The Integration Stage is where the infor- The Reflection Stage is when the user re- The Action Stage is when people choose
people’s motivation for tracking, identifica- people collect information about them- mation collected in the Collection stage flects on their personal information. Users what they are going to do with their new-
tion of what information to track, and se- selves, such as their inner thoughts, be- are prepared, combined, and tranformed may reflect on the information immediately found understanding of themselves.
lection of what tools to use for tracking. havior, social interactions, and their imme- for user to reflect on. after recording (short-term) or after several
diate environment. days or weeks (long-term).
Barriers
Preparation Barriers can occur when Collection Barriers Integration Barriers Reflection Barriers Action Barriers can occur among the dif-
users choose the wrong information to Tool “not having ready access to a com- Transcribing “It'd be neat if I could graph it Lack of Time “Having time to go through every- ferent ways that systems support this
track or when users select an inappropri- puter at the time symptoms happen” straight from the website instead of thing, but that is also one of my big- stage. Some may tailor their behaviors to
Data manually typing in the data to a gest pleasures is finding that time.”
ate tool that does not satisfy their informa- Remembering “Forgetting to record it. Because I spreadsheet” match their goals. Some systems alert
tion needs. These incorrect selections can am often not at my personal com- Visualization “It's hard to get a holistic view of the
users when particular thresholds are met.
puter.” Organization “Collecting is simple. Organizing it data since the time filters are at most
lead to data loss and wasted time. takes some time.” one month and I'd like to look at sev- Some systems provide incentives to moti-
Lack of Time “Not difficult, time consuming at eral months at once.”
vate users to take action. Most systems do
times.” Scattered “A bit cumbersome going to so
Visualizations many different sites [for visualiza- Self-criticism “It's extremely difficult (psycholo- not have specific suggestions on what to
Finding Data “Sometimes life isn’t interesting tions] gically) to look back on my earliest
enough to make me want to write it journals. Much of that information is do next, which is a barrier to applying un-
down, other times I can’t find any Multiple “Difficult to keep organized because very emotional and innocent.”
derstanding of personal information.
worthy writing material.” sometimes data are kept in separate
Inputs places.” Interpretation “Sometimes its very difficult to inter-
Accuracy “Guesstimating mass of food pret the media.”
matching homemade or restaurant
foods against database entries” Search “Not too tough. sometimes have to
wait while search occurs... but it's a
Motivation “Keeping up the motivation to do so, couple minutes at most.”
finding payback for the investment
of time and effort.” No Context “Not having an overlay of changes in
circumstance.”
Sparse Data “Not enough; My collection of data
has been intermittent enough that I
don't get good time series.”
Data is Not “It's really not very useful and it's
kind of annoying. I mean, I walk a lot.
Useful What else do I really want to know?”
Properties
Barriers cascade. Stages are iterative. User- vs. System-driven Facets
Problems in earlier stages affect the later stages. For Stages are iterative; users will incorporate new data, Each stage can be classified as user-driven, system- This property concerns facets of a person's life. Most
example, not selecting the right tool during the Prepa- tools, and processes as they progress through the driven, or a combination of both. In a user-driven systems are uni-faceted, collecting only one facet of a
ration stage may lead to reflecting on incorrect data. stages. For example, a user may change the types of stage, the user is responsible for the activity in the person's life (e.g., Mint for financial matters, Nike+ for
Another example is that problems in the Collection exercises she performs. These changes may require stage, while in a system-driven stage, the system is. physical activity). Some systems are multi-faceted,
stage may lead to sparse data, which may be insuffi- new tools, new types of data, and different visualiza- For example, a user-driven Collection stage may re- collecting multiple facets of a person's life (e.g.,
cient for insightful reflection. tions. Often times, the user cannot bring their old quire users to record information into a spreadsheet, Daytum, your.flowingdata). However, such systems
data along with them. This causes problems because while a system-driven stage may use sensors to track usually present multiple facets in separate visualiza-
it makes comparing between different types of exer- personal information. tions. Many participants expressed their desire to see
cise more difficult. associations between different facets of their lives.
Design holistically. Maintain flexibility. Balance automation and user control. Associate multiple facets.
This property suggests that the development of per- This property suggests that systems should be flex- This property suggests that there are opportunities to This property opens several opportunities to explore
sonal informatics systems should be approached ho- ible to support users' changing information needs. alleviate the demands on the user using automation; how applications can better support awareness of
listically. Of course, we should take inspiration from Some examples are support for easy importing and however, developers should consider the tradeoffs associations between different facets of life.
different fields to resolve problems within each stage exporting of data and rapid iteration so that users can (e.g., inaccuracy of automated tracking and loss of
(e.g., visualization techniques from the information vi- hone in on the questions they want to answer. user control.)
sualization community), but development should not
focus only on one stage, but consider the whole ex-
perience of the user throughout the different stages.
COLOPHON
Font
Helvetica Neue