Design Thinking: Data Intelligence

Yesterday big data was all the rage; today we have intelligent machines and processes to
leverage that data wisely. Learn how modern data intelligence helps designers provide a more
personal and human design experience. Randall Elliott, associate director of design
technology at frog, shows how to use data in creative problem-solving: to generate ideas,
create better designs, and connect with users in meaningful ways. He tackles data generated
by computers as well as that gathered through interviews and other hands-on means. He
shows how to collect, analyze, and visualize the information to help solve your customer's
challenges, leveraging intelligent tools and agile processes such as rapid prototyping.

Learning Objectives:

 Design intelligence best practices

 Avoiding data paralysis

 Applying data intelligence to design

 Partners in design intelligence

 Visualizing data intelligence

Instructor
Introduction
Welcome (Viewed)
Welcome

- I'm Randall Elliott, an associate directorof design technology at Frog, a global digital
transformation and strategy firmthat transforms businesses through design and
innovation that advances the human experience. At Frog, we've used design
approaches throughout our 45-year history to discover what makes meaningful connections
with people. This course will share how we use data intelligence to create more human
experiences. Whether it's data collected from speaking to people or captured from computers
and coalesced by data scientists, these are core inputs to creative problem-solving. This is
often called design thinking. Businesses around the world are using this to open new
marketsand create competitive advantage. In this course you'll learn how data intelligencecan
generate ideas, solve complex problems, and connect with people in meaningful ways. While
this course won't cover in-depth data science instruction, it will identify how to integrate data
intelligence into your teams as well as how to approach and explore the technology that
makes it intelligent. If you've heard about how data intelligence creates new experiences, or
read about companies who are using it to explore new markets and wondered how you could
bringmeaningful experiences to people with it,then this course is for you. Plus we've added
an interactive feature to the courseenabling you to collaborate with othersand help data
intelligence inform your design decisions. Okay, let's go.
1. A Foundation for Design Intelligence
How design intelligence provides value
How design intelligence provides value

- For a moment, imagine what it was like when we had little data, and therefore, could not
offer much information to the people who used our products and services. Here's an example.
Remember when an electronic device simply told us when it was about to run out of power?
Sometimes we were told what percentage of battery was left, but that was it. Today, we're
notified how many movies we can watch, or ways to save power and extend use. Did you
catch the distinction between those? In the past, we were provided with very little helpful
information, almost nothing about how much time was left or what we could expect from it,
while today we receive clues to help us experience more from the product. Let's take a
moment and take a deeper look at the evolution of these examples. When we have valuable
data, we're able to make it more valuable by making some inferences. Simply put, an
inference is a conclusion reached on the basis of evidence and reasoning. This data enables us
to design the use of these inferences in a way that can be communicated clearly to our
customers. In this case, we estimated how much power was left, and then inferred what you
may be able to do with it, for how long and how you might save power. To make the value
even clearer, what if we had simply displayed all the raw data, like let them know that their
battery, for example, is 1.5 volts, and maybe 1.6 when full, and 1.1 when almost empty, then
told them the rate of change of their battery in volts. Do you see how those values on their
own did little to impact the person's experience? Then when we apply design thinking and
user-centered design, we turn those values into something immediately tangible. The way that
we chose to communicate those inferences regarding battery life changed the human
experience. It gave us more capabilities through intelligence. The data intelligence was there,
and when we applied design thinking, that data became even more valuable. That synergy
between data and design and its output, we call design intelligence.

Design intelligence best practices

- We have many design methodologies and frameworks we've developed. From designing
forums to techniques that reduce the cognitive load on a person, in Design, we don't have a
shortcoming when it comes to best practices for just about anything. Except for data
intelligence. Do a quick Google search on data science or data design, and you will mostly
see database, data structure, data visualization, and voice interaction best practices. There are
a few gems here and there though. So check out the course links and book recommendations.
But when you sit down with your team to create, and you'd like to leverage data intelligence,
where do you start? We really need a framework that connects the needs of the people we're
designing for to the valuable data intelligence we have. Here's a framework to use during
ideation and concept creation, when you'd like to consider how using data intelligence may
create even better solutions. Instead of getting wrapped up in the details and specifics with
actual models, algorithms, or technologies, we'll use this framework that is focused on
helping the people who use your stuff. So this framework helps describe how machine
intelligence enables. From the perspective of someone's experience, rather than focusing on
the literal technology, or algorithms. It consist of four enablers, or answers to the person or
business need. Humanize, expedite, transform, and discover. Let's break these down one by
one and talk about how needs of people translate into each of these. If you hear from people
that something you currently provide or represent in a concept, is hard to decipher, you might
try humanizing the experience to further clarify it. How might we humanize a machine-based
interaction? The most common application of this solution wee see today is personalization.
For example, think about how someone you have known for years is typically easy to talk to,
and understands your needs. This is because they personalize the output so you have more
context. In the near future, we will have even a wider history of recorded experiences to
design around. Okay, let's move on to the next design intelligence enabler. Expedite. Are you
looking to free up time for your customers or the business so they can intervene in tasks that
technology is not yet ready to handle? Perhaps they want to be part of those interactions
because of the human value they bring to it, or the emotional value it delivers them.
Leveraging data intelligence to expedite or replace rote, repetitive tasks, enables machines to
free up humans to deliver their best value to your business. Consider this possible design
intelligence enabler whenever your customers could use some time to perform more valuable
tasks. Alright, now onto the next enabler. Transform. When framing a design challenge
around an organization, the people and its values were often looking for opportunities to
transform the business. In this case, transform data intelligence you have on market
opportunities or customer behavior into an input used to frame your organization's goals, and
how you might achieve those. For our final enabler, discover, we can use any patters of needs
uncovered by data intelligence to reveal new ways to make the human experience a better
one. For example, if a wholly, new need is discovered, then it can be used to develop a new
product or service, or they can identify the need for a new design solution. To get on a very
personal level with your customers, these discoveries could proactively help those who may
be in need of assistance before they reach out or complain. Lastly, these could guide them as
they use your products. So they have less points of resistance. So the next time you have
some valuable inputs for machine intelligence, design intelligently by asking yourself, if any
one of these four enablers might drive a design that fulfills the need of your customers in a
new or previously unimaginable way.

Design research and data intelligence

- Okay, you and your design team are about to hit the field for research that will include
observations from customers framed by those you've made by interviewing stakeholders.
You've also been blessed with a data scientist to offer some new perspectives on your
customers, product, and business. How do you incorporate that data intelligence into your
insights and ideas? How do you transform these perhaps disparate data points into a concept
that you can rally around? To help transform those ideas, create something that serves as the
singular source of truth, that way the data intelligence you have becomes part of what feeds
into your ideas. When you succeed at this you'll find that more will become the hero of a
concept. Another great value of doing this is that you can demonstrate an evidence based
guess. That's right, it means just what it sounds like, you have data that serves as evidence for
your design hypothesis. In a nutshell, if you have a hunch or even an observation, there may
be data intelligence to back it up. Potentially making it more valuable or providing necessary
proof for that skeptical stakeholder. But that means you must directly incorporate those inputs
along with your observations during synthesis. And how do you do that? Systemize your
source collection and insight development. That means make sure that along with analog
recording methods, everything is systematically recorded digitally. This could be a
spreadsheet that everyone uses during research like the one here. Every column should have a
selection where possible so when entering unstructured or freeform data every bit is
categorized in the same way. Also note how each one is coded so everything may be cross
referenced. These codes are often best designed to capture important references such as
geography, customer, company, or even market segment. Imagine what happens when you
incorporate something like a standard spreadsheet into every bit of research you conduct. One
immediate benefit I've noticed is we can corelate what the stakeholders feel the customer
needs with observations we've made meeting with customers. Before we go any further it's
important that we talk about ongoing data conversion. You see, as a group you'll likely still
leverage analog recordings. You know, post its and worksheets. So an individual or a group
will need to be responsible for recording these insights. They in turn become the field
recorder. Now that you have a system to collect your data digitally, you can leverage your
data scientists and statistical programmers to cluster the data. We're not going to go deep here
as this is so dependent on the project and could be a course in and of itself but we will touch
on what this means and the value it can provide. In a nutshell you can use algorithms to group
observations and you can also cluster these to create insights. These can then be compared
and contrasted to the insights you've come up with. You can get really fancy, paired into
those alike and compared to the norm to help eliminate any bias that may be unknowingly
introduced. Last but not least, this is a fantastic way to verify what may be most valuable to
research and experiment further. Next time you're putting a research plan together make sure
you find a way to create a structured digital artifact of your observations. Record what you
find during synthesis and create a single source where these all live together. Oh, and make
sure other findings are recorded there too throughout your process. These strategies to
digitize your observations and insights will allow you to uncover more data intelligence.

How to avoid data paralysis

- The digitization of our world has brought with it the ability to track nearly every action of
our customers. This has provided us with more feedback and data than ever. Have you ever
sought feedback on a solution you and your team were working on? You likely did in hopes
that it would help you evolve or validate it. We often find that the most valuable feedback is
able to provide us with clear direction. But what happens if we receive so much feedback that
there are many points of view? What if they are at conflict with one another? It can be
downright overwhelming. It's because we have too many directions or so much feedback that
it actually makes it less clear. In contrast, if we receive only one form of input, we often feel
like we haven't received enough feedback to derive a sense of direction. Today, it can be easy
to look at the data we collect, such as A/B testing, being the only or most important form of
feedback. It's become easy, so sometimes we don't scrutinize it or seek feedback from other
sources. Worse yet, sometimes we see it as the end of the test and the direct input to the next
instead of the beginning of another concept or experiment. What do I mean by experiment?
Testing concepts used to be viewed almost exclusively as an experiment. We'd test it, look at
the results, and attempt to infer the outcome. Then we'd create a new one and test again.
Unfortunately, with data today, we often see what used to be deemed experiments used as
exclusive tests that do not seek alternative explanations or solutions from other research. This
could be hard to grasp at first, so let's use an example. Imagine we work for a huge online
shopping site, such as Amazon, and our design team has been told that we need to find ways
to help people buy 4K televisions easier. We learn that this new ask is largely because they
have some recent data showing that customers looking to purchase these fancy new TVs have
an extremely low rate of completing their purchases. We could start by making some
educated guesses as to why they are having trouble completing their purchase, devise several
alternative ways to purchase, and pick two to test. Then, based on the results of the test, select
the best one and create two modified versions to test again until one shows significant
increase in purchases. That sounds logical, but what if perhaps the failure occurred in our test
time frame, unknown human behavior, or execution? We'd likely not know that from our A
or B tests alone, which means we can't say for sure one performed better than the other or that
either would really help the person purchase a TV. The bottom line is when we look at it so
narrowly, it seems like a good way to judge experiences and reduce the risk of failure to the
business. The problem with this approach is that it absolutely ignores all kinds of other
inputs. I call that data paralysis. Instead, we should start by looking at customer complaints.
Talk to people interested in purchasing those TVs. Develop a few ideas in addition to running
some tests. Walk some of those folks through our concepts. That way, we can see in here
what helps people buy these fancy new TVs. Now we have a sense of direction and data to
inform our designs. Don't forget the data is still valuable. When coupled with design research,
it can reduce the amount of people needed in your study, thereby reducing costs and
potentially enabling you to go deeper with your contextual studies. That could also lead to
uncovering new research and the people whom you should interview. I don't know about you,
but I feel like, in the past, when these new ways to provide us feedback emerged, we focused
too much on them alone, potentially missing many opportunities to truly transform the human
experience. So let's move forward and continue to seek more feedback from data, and
leverage it to infer how we may better answer our customers' needs. Go forth and experiment.

Add intelligence to your concepts

- Okay, so you have everything recorded, including digital versions of your observations,
hypotheses, and insights. Now how can you include data intelligence in your ideation and
concept development? In the room, you will be ideating and creating concepts. Write these
four questions in a prominent place. Do you nee to provide more clarity? Do you need to
provide or save more time? Do you need to better frame customer challenges or
organizational goals? Do you need to provide more or proactive guidance? If these sound
familiar, they are questions derived from the four design intelligence enablers, humanize,
expedite, transform, and discover. As you create your early ideas, all those initial seed ideas
that contribute to your selected concepts, use these techniques to leverage data intelligence.
The first is the pivot. After you come up with a concept, or after someone else shares theirs
out, ask yourself if it is solving any of those four questions. If it does, take each one it
answers and pivot it into a concept that solves it by using data intelligence. For example, let's
say a teammate created a concept to reduce the time it takes for a customer to check out.
Using the pivot, you'd ask yourself, instead of reducing the check out time, how might I
provide them with more time to shop using data intelligence? Worksheets can serve as a great
analog recording device and a fantastic method to note what data points you've folded in, so
make sure to download the worksheets available in the exercise files section. Please pause the
video if you have not already downloaded the data intelligence concept worksheet. For the
best experience, you'll want to print a few out and follow along. Okay, so now that we have
the worksheet and something to write with, let's walk through it together. Oh, and if you're
already smack-dab in the middle of a project, feel free to follow along and use your
observations from your research to try it out. Use this worksheet to prompt wholly new
concepts. Keep it focused on the person by pulling needs and pain points from your research.
This is why the worksheet starts with identifying one of those as a foundation for your
concept. Okay, so now you see why the next step of the worksheet asks, what value can we
provide the person to? Followed by values from the four technology enablers. Save them
time, help them understand, guide them, or make them smarter. Use as many of these as you
like. If you'd like a greater breadth of ideas, try running each need or pain point through a
worksheet to develop many disparate ideas rapidly. On the other hand, if you want more
depth, try taking the same ones another direction until you've absolutely exhausted your
ideas. You'll find that by adding intelligence with human needs in mind to your concepts,
you'll create more valuable experiences. Let me know how you use the worksheet and any
modifications or tricks you may have added. Best of luck.
2. Applying Data Intelligence to Design
Data science and design
Data science and design

- With the popularity of big data many companies have amassed an unbelievable amount of
data around their businesses and their customers. In the last five years data analytics and
science have ushered in ways to leverage all of this data as companies look for new ways to
use it. This quest for deriving value from the data has led to the proliferation of artificial
intelligence and its strategies such as machine learning. So now many companies are looking
to leverage the data they've gathered by designing innovative customer experiences. Let's
pause a moment and define some of these new terms. Big data refers to extremely large data
sets that may be analyzed computationally to reveal patterns, trends, and associations,
especially relating to human behavior and interactions. Artificial intelligence generally refers
to machines that can think in some ways like humans. Machine learning is commonly
associated with AI and is a field of computer science that gives computers the ability to learn
without being explicitly programmed. Data science applies the scientific method to data as a
means to infer valuable evidence that can be used to form new ideas. Now that that's all out
of the way, let's talk about how data scientists and designers think and work. In general,
scientists use a methodology called the scientific method to solve problems, they study an
organism or an organization, they form a hypothesis, test it, and record the results. Often
those results lead to a new one and the process is repeated. Designers use a variety of
methodologies and apply processes that frame the problem and develop a point of view learn
about people to develop deep insights and make a solution to test and learn more. Did you
catch those similarities? The two approach problems and seek solutions in much the same
way, now that's great and everything but just because they're similar doesn't necessarily mean
they're complementary. Well think about it this way, quite often when we arrive at a solution
based on research we're still working with just a hutch, while we have a plethora of user
validation and testing methodologies, they often fall short it's a delicate balance. The people
that use our products provide the feedback, this provides much data, and must continue to do
so for us to provide the best experience. So you see their input provides the solution while
we've got to have data science to develop the mechanisms to accomplish that. To illustrate
this take a moment and think of all the recommendation services we see today.
Recommendation services could be suggesting the next movie to watch or an advertisement
customized for your needs or shopping preferences. There are hundreds and perhaps
thousands of (mumbles) of these today. Now think of the ones that actually provide real value
to you, the ones that make your life easier or tell your friends how cool it is. You can only
think of just a few right? Okay so how do data scientists and designers actually complement
each other? I'll let you in on a little secret, the successful ones utilize a well-crafted
orchestration of a design that creates less friction while collecting data that can make the
experience even more personalized. Make sure you lay the foundation for collaboration
between design and data science on each project. We find that by planning from the
beginning parallel tracks of work with shared collaboration moments along with milestones
to be most effective. Designers and data scientists approach challenges in a similar fashion.
When coupled with shared tasks and goals they create beautiful experiences that make our
lives better.

Rapid prototyping with data

- It may seem like these new technologies are too complex to rapidly prototype, but you'd be
wrong. If you assemble the right team of multi-disciplinary designers, everything is possible.
There are two approaches to rapidly prototype with data by co-designing with a data scientist.
The first is using a data-driven prototyping approach. Those let you import real data into your
web and mobile prototypes, and build dynamic data lists and data grids within charts, menus,
and other data simulations that actually behave like the real thing. The second approach is
through data prototypes. These act as descriptions of a technique either through a proof of
concept or just an example of how raw data has been transformed into something meaningful
to your users. For the first approach, data-driven, you'll need to collaborate early on, while
you're both performing design and data research. That way, you'll be able to integrate real
data from the data scientists research. Then, you and your data partner will be able to discuss
which data-driven prototyping tool best supports your needs. This synergy of collaborative
research and data-driven prototyping guarantees you'll bring some of your best concepts
forward. A few tools you want to consider for data-driven prototypes are Justinmind,
Envision, and Azure. We'll not be diving into these during this course. Rather, we'll be
providing you with frameworks along with skill and team-building strategies. If you want to
drive into these after you complete the course, you're in the right place. We have a number of
great courses covering those applications right here. To use the second method, the data
prototypes approach, data prototypes are much like design prototypes. Just like how you
sketch an experience, you can work with your data scientists to mock-up the result of their
work. This can be an example of how the raw, or original data looks, and then what it
becomes when they transform it into a valuable insight. Using these data prototypes, you can
then articulate how you would leverage them, how they would be communicated to your
customers, and to your business. These prototypes can range from brief descriptions of how
you would transform the data, and a presentation communicating the value it provides
through concept sketches, examples pulled from another industry, or even books or movies
that have inspired the creation. Other times, you may need to work with a technologist to use
fake data provided by the data scientist and integrated into a basic application to act as a
proof of concept. Selecting which technologies are appropriate is an entire course in itself.
But suffice it to say, we can build real-looking digital products very rapidly today, with
motivated teams who have a diverse skill set of technology and design. Okay, we have two
branches of rapid prototyping with data. Data-driven prototyping, and data prototypes. So
which one do we use? Choosing which one will most likely have to do with time and the skill
set available. If you need to move fast and you've got a small team of a designer and a data
scientist, go with data-driven prototyping. If you have more time and have designers who can
work conceptually with science and technology, go the data prototyping route.

Framing with data intelligence

- Solving problems through design has developed into many methodologies over the last
several decades. One of the most prominent methodologies is design thinking. Design
thinking is a method for practical, creative resolution of problems. It can be used to solve just
about any kind of problem. It's a form of solution-based thinking with the intent of producing
a constructive, future result. Often, it's human-centered, meaning it focuses on who it will be
providing the solution for. A key to this creative problem-solving is framing. Framing simply
means accurately identifying the challenge and who needs your help. By identifying the
problem, we can focus our ideation only on the most important needs of the people using our
product. Data intelligence is a fairly new input to design. Often it comes from more of a
science methodology, and it starts with a hypothesis. One thing to note, it can be helpful in
many situations to treat data intelligence like an organizational specialty or technology that
gives an edge to the business, rather than getting too caught up in the fact that it comes from
data intelligence. Another way you can leverage data intelligence is by incorporating it within
your design thinking strategies. Since design thinking uses both analysis and synthesis,
design intelligence can help frame the problem very well. So once we frame it, how can it
actually help solve the problem? To explore many possibilities we use divergent thinking and
use convergent thinking to narrow and select those that solve the challenge best. Injecting
data intelligence at this stage can be super helpful in two ways. You can look for ways to
solve the problem using data intelligence to come up with potentially wholly new ideas, or
you may use it to seek ways that bolsters existing ideas as a way to narrow the solutions you
further design. As you can see, there are many ways that design intelligence can help us
frame the problem, and it can help us choose the best solution.

Intelligence will transform designers

- Many people are once again concerned about machines taking their jobs. And for many
designers it's now their time to worry. This stems from a variety of recent machine-learning
techniques that automate some of what used to be decisions made by human designers. We've
seen this especially within marketing and advertising, due to several recent innovations that
customize the ads and target who receives them using artificial intelligence. In addition to
this, big companies, such as Google, are experimenting with computer-generated art. So it's
not surprising that many people have jumped to the conclusion that it won't be long until
computers can develop designs for ads and products. At the same time, Harvard and many
other prestigious institutions are encouraging several vocations that in the past weren't
associated with needing specialized design skills to do just that. In fact, in a recent Harvard
Business article, the number four recommendation for managers was to work like a designer.
They even went on to say, While managers' own creative abilities are vital, perhaps even
more important is their ability to harness others' creativity. As you see, it really depends on
the lens you look through. It can be very the sky is falling or incredibly futuristic. Personally,
I prefer the opportunistic and the futuristic.
3. Our Transformation through Intelligence
Great experiences from predictive data
Great experiences from predictive data

- In the past, we tended to hear only about the data marketing snafus that focused on invading
our privacy through data. These were largely added marketing-focused. Today, companies
and researchers are using techniques like machine learning to predict minute behavior
patterns. Some are even using these to intelligently create extremely complex digital
personas that can then be used to model and predict what they might do or say. This might be
hard to grasp or imagine at first, so let's start with a well-known example. Facebook
predictive political analysis. You likely have heard about this, and might even looked your
own up if you use Facebook. By analyzing millions of data points,Facebook has developed a
system that can predict where you are in the political spectrum and how politically
active. Think about that for a moment. Let it set in. We have the means to understand if
someone is liberal or very liberal, or how likely they would be to engage politically. A
person's politics can be incredibly complex,something that is tied into their belief
systems, morals, culture, and even religious views. We can accomplish the above by
analyzing their behaviors,making a model, and then testing itagainst millions of other similar
data points and people. As you can now imagine, many are rapidly experimenting with these
technologies to literally drive a better user experience for their customers and employees. For
example, cybersecurity has been a big issue recentlywith data breaches happening nearly
every month. Securing this data has become big business, and it's difficult to do well. So,
institutions are experimenting with using behavior analysis to create predictive personas in
order to prevent people from doing something maliciousbefore it happens. Think of it like an
early warning system for weather. It analyzes user behavior on the networks and compares it
to models to predict what they might do or not do. How might you use predictive analysis to
improve your product's design? If you stumble a bit with this, check out the link to the
worksheetincluded with this course. It is designed to help you unblock and unlock new ideas
around using design intelligence to predict. Perhaps you want to know what your customers
need before they have to ask. Or, maybe you want to be able tointroduce them to other
services that people like them found the most useful.Give it a shot and let me know in the
course comments your favorite ideas, and how you executed them. Please share failures and
success stories.
A new skill set for data intelligence
A new skill set for data intelligence

- We've all heard by now how big data, machine learning, and artificial intelligence are
impacting our world. What we may not realize is that this affects how we design now
products and services today. It's something we should be prepared for today by acquiring a
new skillset. Ironically, it's like not thinking about how your company may be interacted with
using a smartphone. Technology like AI is such a moving target that it is hard to see
sometimes. After all, AI is often defined as something not before achievable by a machine,
which means, once a machine can, it's technically no longer considered intelligent. Once
again, it's just a machine following a program. Understanding these new technologies,
techniques, and inputs, can frame our design challenges today and future-proof them. So what
are the most valuable and fastest ways you can become more skilled at leveraging data
intelligence? Future designers' greatness will likely be judged by how well they train the
intelligence systems to interact with us. They will have to design them so others are able to
train them as well. That shift in the focus of design will change the necessary skills needed.
First, thanks to intelligent machines, people will be free to apply design thinking even more
broadly. Second, more people will be needed to design how we interact with the products and
services of tomorrow. Lastly, most people will need an understanding of how intelligent
systems work so they have the knowledge to create new ideas. If you don't consider yourself
a designer, either by occupation or education, you should acquire design thinking skills.
There are several great integrated design programs, if formal education is an option, such as
the Integrated Design Program at MIT or the University at Texas. However, if formal
education is not an option, look into those design programs' curriculum to see what
knowledge areas and skills are covered, then research them online and in books on your own
time. You can gain much knowledge on your own if you're motivated to seek it out and study.
Also, there are several online and fast-paced programs that teach design thinking. A great
start is by watching the other courses in the Design Thinking series right here. If you're a
designer, add technology and psychology skills to your repertoire. Learning some basic
Python data science skills today is quite simple and rewarding. It will make you a valuable
partner to a data scientist colleague. Several companies that provide data intelligence services
like Google and Microsoft also provide simple introductions to using their service and will
have you up and running quickly. If you're unable to take a few psychology courses, consider
reading some of the books pertaining to AI and psychology listed in my course
recommendations. Whether you are a designer or a CEO, the future will require design
thinking skills and intelligent systems knowledge. It will be our responsibility to design these
to treat us as we would want to be treated. It will also be a great opportunity to shape how we
interact with the intelligent products of tomorrow.

Partners in design intelligence

- So you'd like to design intelligent things, huh? Well, you just need these three skills:
programming, data science, and business knowledge. That's all, you say? Kidding aside, what
if you don't have all of those? Perhaps you have never programmed, or it has been many
years in a language long ago. Maybe you know about statistics and analytics, but lack the
skills to perform those duties. Nothing to fear, my friends. If you desire to work with artificial
intelligence, to design intelligent products, you simply need the right partners. Just like any
business, having the right partners can make the difference between failure and success. So
let's start by talking a bit about the kinds of people who are able to use AI to design
intelligent products and services. First, the programmers. These folks will either be statistical
programmers or likely have the word data prefixed to their titles, such as data analyst or data
processor. Next, the data scientists. This is a relatively new practice that usually involves
someone who has received traditional teaching in how to apply the scientific method and
might have focused in a particular area of science. In addition to their science teachings, they
have also focused on data research, analytics, and statistics. Often, they also have
programming experience, as well. You'll need folks with business knowledge, too. They are
often business strategists, or those who have focused in business whether it be formal
education, job experience, or entrepreneurship. Lastly, in some or all of the categories, you'll
need people with mathematical knowledge. Having people with this knowledge is
foundational. The sum of your new partner's expertise should include probabilities, statistics,
linear algebra, and mathematical optimization. The whole point of this is that it's okay if you
don't even really know what those are. Surround yourself with team members who do. In
addition to new skills, you will likely need new partners that have skills you are unable to
rapidly acquire. These new partners should cover these three areas: programming, data
science, and business knowledge. In addition to these, they should have a good mathematical
foundation for working with artificial intelligence and especially concepts like machine
learning. For more detailed information, check out the course links and let me know how you
use this to build a team for the future.

How design disciplines must evolve

- Designing with intelligent systems whether they simply provide data that alters content for
site visitors or can actually design themselves, requires each discipline to better support the
people that interact with them by evolving their craft. As a designer, we are all likely already
working with data intelligence in some way. After all it's impossible to ignore. So it comes to
no surprise that we want to know how to learn and do more. In this new world as designers
we are responsible for creating a safe and natural interaction with these machines. Whether
we are teaching them or they are learning from the people that utilize our products or
services, visual designers who design DLS design language systems will have to evolve their
systems they need to support dynamic systems. While interaction designers will deepen their
understanding of human interactions with machines, as teachers and guides. For UX and UI
designers, there are opportunities to evolve how you design the experience and how the user
interfaces can best support design intelligence. For the technologist especially design
technologist, gaining knowledge of the data and technologies that enable designing with
intelligence can evolve your discipline rapidly. Couple that with learning how to prototype
intelligence systems and you have a huge opportunity to play an important role in the new
designs. Any designer or technologist will stand to prosper greatly by evolving their craft to
help design with intelligence. For most it simply involves flexing those research muscles and
learning as much as they can about the technologies and possibilities. It truly is an exciting
time for design and our design practices.
4. Seeing Data Intelligence
Visualizing data intelligence
Visualizing data intelligence

- What happens when the model or data produces a value? Is the output of an algorithm
inherently valuable? The answer to both is no. It's all about the delivery of that
output. Central to our charter at Frogis improving the human experience through design. Big
data is well, big, and everyone is capturing oceans of it in hopes of gaining new insights
through intelligence. But how does this make meaningful connections to people?Perhaps
Cecilia Aragon, director of the Human Centered Data Science Lab at the University of
Washington said it best by posing the question how can we be sure not to lose the compelling
and inspiring stories of individuals in the sea of aggregated data at scale? Data science can
provide this intelligence, creating new connections to the people whomexperience your
design. We've been using charts, graphs and visualizations for a long time to tell a
story. Perhaps even longer than you'd imagine. In fact, one of the first data visualizations is
sunspots over time by Scheiner is 1625. And the first time for a new visualization
technique. That technique was later called the principle of small multiples by Tufte. What
Scheiner did even way back then was order the data in a way that illustrated the small
changes. This is very similar to how we transform data using algorithms, so we may be better
tell a story with that data.This would totally make sense if we could simply look at and play
with a data visualization in real time. It would be so cool if we had a tool to do that. In fact,
I've made such a tool. This is using data from logged observations of interactions within a
dolphin community in New Zealand. They identified and named each dolphin, then every
time a dolphin had an interaction with another, they logged it.Here's what that data looks
like when transformed into a format called JSON. It's just a bunch of data, and you can see all
of these interactions and they're not very organized. But now look what happenswhen we use
algorithms to transform that data visually. For example, you can see now, even with this first
algorithm, a little bit of clustering. It's not very helpful, and it's pretty close to what the raw
data probably looks like to your brain. But when we actually apply just a simple
algorithm called breadth first, we can actually instantly see a transformation in the data. We
can see that one dolphinlooks like a really popular dolphin and perhaps is connected in every
way to all the other dolphins in the community. Now if we apply another algorithm that's a
circular algorithm, we get another view of the same data. Because of how the algorithm
here is showing us maybe clusters of areas around this circle where it's denser than
others, which might mean that many parts of the communities, rightare more active than
maybe some of these other areas that are quite sparse. Then we can apply yet another
algorithm that starts to use some clustering techniques. So now we actually see the sort of
little communities within the communities, and we can start to infer that maybe this is a
group of dolphins that have known each other for a long time. Maybe they're family. Maybe
they have a role to play in the community. Maybe they protect the community or they're
lookouts. And this data really starts to come to life. What this means to your business and
customers is think for a moment that this data isn't about a dolphin community, it's about all
of your customers and the communities within your customers that use your products and
services. Think about how that could be applied to this. Think of what you could learn about
your customers and your business by looking at the data in new ways using these visual
algorithms.