Customer Behavior Analysis with SQL and Tableau

Section 01: Introduction

Section 02: Customer Engagement Analsis Case Study

Case Study
Why is it important to study customer engagement? We live in a competitive environment with an
abundance of choices. If a customer is unsatisfied with a given product and doesn't find it intuitive or
straightforward, they could easily find an alternative solution that works better. Moreover, customers
who engage with a product to a greater extent are more likely to stay loyal for a more extended period,
which is crucial, especially regarding subscription services. What keeps us motivated at 365 data science
is when our students engage with the courses we create and most importantly, gain valuable
knowledge. Imagine William, who's just activated his trial period to a movie streaming platform. Upon
scrolling through the movie catalog, he realizes it's difficult to find the information he's usually
interested that is starring actors, movie, resume, origin and rating. Moreover, each page loads slowly.
Even though his Internet connection is stable. To top it all off, after playing a movie or two, we'll find
that the quality of the stream doesn't meet his expectations. Unfortunately, at the end of the trial
period, he decides to spend his money on an alternative provider. To reduce such situations to a
minimum and pinpoint potential areas of improvement. It's essential to know how your customers
experience your product, how much they engage with it, and if the retention decreases with time. That's
precisely what we'll do with the 365 data science platform. If you're viewing this course, you're probably
already familiar with the product, But let's do a quick overview. 365. Data Science is an online learning
platform that aims to deliver affordable, high quality courses on topics relatable to contemporary data
scientists. Quizzes, practice exams and a course exam complement each course on the platform. Passing
a course exam grants students a course certificate, a valuable addition to a professional CV
demonstrating their proficiency in a given subject. To perform analysis on student engagement. We
should understand both the platform features and the terminology we use at 365. Every company
creates its own vocabulary and definitions throughout its development. For example, what is considered
an engaged user on the 365 platform might differ from what a social media platform would convey.
Other terms whose definitions vary from company to company could be stale, resurrected, retained or
churned users. For now, let's present the in-house definition of the terms you'll work with throughout
this course. The first term is engagement defined as beginning one of the following activities a lecture
from a course, a quiz or an exam. Engagement can be recorded in different time scales. Later in the
course, we'll create a visualization showing daily and monthly student engagement. The second term in
three six fives. Vocabulary we need to define is onboarding describing a first time engagement. Imagine I
registered on the platform on January 1st, then closed my laptop and continued my day. Then on
January 2nd, I began learning by enrolling in the introduction to data and data science course and
watching the first couple of lectures. I'm therefore considered onboarded and an engaged student on
January 2nd. Finally, we'll discuss free plan students and paying ones during the course. Since the 365
platform is a subscription based product, intuitively, the former type of students refers to ones without
a paid subscription, while the latter refers to students with a paid plan. The types of subscriptions a
student can purchase are monthly, quarterly or annual. All right. Now that you have a better
understanding of the platform's inner workings, we can pose the questions we want to answer by the
end of this course, including. How engaged are the students inside the platform and how can this metric
be improved? How long do students stay engaged on the platform and how can this period be
extended? What's the difference in behavior between free and paid students? Which are the most
popular courses on the platform? How many students sit an exam? In the following lectures, we'll
formulate these questions more carefully, giving us a stepping stone to thinking about the visualizations
that would later enter the Tableau dashboard.

Define the key questions (Part 1)

Planning is the most time consuming part of any project as you're creating a product entirely from
scratch. It's not something that can be searched on the Internet or looked up on an answer sheet at the
back of the textbook. Before implementing a big project that would take a lot of time to execute, such as
creating a multi-page dashboard. It's good to have it scrutinized, discussed with colleagues in detail and
carefully designed on paper. Creating a good action plan is worth your time as it's a preventive measure
against repeating the project. Or even worse. Start the project again from scratch. So before creating
charts in Tableau, let's first write down a list of meaningful questions that the dashboard needs to
answer. For each question, let's also try and think of a hypothetical outcome that will later be proven or
disproven by analysis of the data. Preparing yourself with certain expectations is essential, as it could
help you spot suspicious behavior and possible mistakes in the analysis. All right, let's go through the
questions. Starting with the following. What is the number of engaged students on the platform? How
does this number differ between free and paid students? How does this number change with time? Is it
affected by marketing campaigns and other events? The main goal of a company is to increase daily
engagement among students. A reasonable expectation is that the number of engaged students who
visit the platform without a paid subscription is significantly higher than those of engaged paying
students. Marketing campaigns also play a significant role in the growth of the number of engaged
students when advertisements are run and subscription prices drop. We can also add what is the
number of onboarded students compared to the number of registered students? This is a vital metric to
consider as it directly refers to the fraction of students who have found the platform intuitive and
intriguing enough to either start a lecture, solve a quiz, or try to pass an exam. Unfortunately, there is no
universal benchmark for what percentage of onboarded students is considered good. Since e-learning
companies define onboarding differently. This number varies between 30 and 50% in some sources, but
we can argue that our definition of an onboarded student doesn't require much effort from the student
side. We don't demand a student have completed a course, an exam or a quiz, but have simply started
it. This measures how easily users navigate the platform, how fast they find what they're searching for,
and whether the platform appeals to them. Therefore, an onboarding rate of 50% or more would be
great to report. If we instead report about 20% or below, this should raise suspicion and we need to
investigate why students aren't eager to engage with the product or whether the right audience has
been targeted. The 365 team is constantly searching for ways to engage newcomers with the content by
employing new platform features, conducting surveys and marketing campaigns, improving the user
journey and experience, suggesting a course to start with and creating new and relevant content. For
example, the team unlocked its platform for free in August 2022, attracting many new students and
expanding its audience. We also released a new gamified platform version on September 16th, 2022.
The expected outcome was to see an increase in student engagement and the onboarding rate during
that time. Later in the course, we'll analyze the data together and see if these expectations were
confirmed. Okay with that cleared up, let's jump to the following questions. How long do students
typically stay engaged on the platform? How does this number differ between free plan students and
paying ones? It's reasonable to suggest that free plan students who never convert to paid ones stay on
the platform for only about a week because there's a limit to the content offered to unpaid users. On
the other hand, paying students would ideally stay engaged throughout their paid subscription, which
can last a month, three months or a full year. The next question is, on average, how much content is
consumed by students in terms of minutes watched? How does this number differ between free plan
students and paying ones? The average number of minutes is the sum of all minutes watched on the
platform divided by the number of students. It's expected that on average, a free plan student would
watch about 30 minutes as this is the typical length of the unlocked sections of the courses. It's not so
easy, however, to estimate the same metric for paid users as each user is interested in different courses
of different lengths and stays a paid 365 member for a different duration. Still, this number is expected
to be significantly higher than the estimate for the free plan users. After all, having paid for a product, a
user would be more eager to get the most out of it than a free plan user. Another question whose
answer could be valuable is the following. What is the free to paid conversion rate based on the minutes
watched on the platform? First, what is a free to paid conversion rate? This metric gives the number of
people who have purchased a subscription as a fraction of those registered on the platform. Presented
in percentages. The question aims to understand the relationship between the minutes watched on the
platform and the subscription rate among students. Our initial hypothesis is that students who haven't
watched any content are unlikely to convert to a paid membership. On the other hand, students who
did engage with the content would decide whether a paid subscription is worth it. Presumably after
completing the free preview of a course, typically between 20 to 30 minutes. Therefore, that's the
amount of content expected to convert most free users. And finally, what is the average subscription
duration based on the minutes watched on the platform? This question aims at checking whether the
correlation between the number of minutes watched on the platform and the subscription duration of
the paid users is linear as is logical to assume. In the case of non-linearity, we can have two student
scenarios try to consume as much content as possible for a short period or learn inconsistently and as a
result, stay subscribed for a long time, but only learn slowly. All right. So far, so good. We're making
significant progress and posing a central questions. In the next lesson, we'll continue asking the
questions we'd like our dashboard to answer.
Sau khi phác thảo các câu hỏi mà chúng tôi muốn bảng điều khiển của mình trả lời, tại sao việc đưa ra
giả thuyết về kết quả của từng câu hỏi và chuẩn bị cho những kỳ vọng cụ thể lại cần thiết?

Kỳ vọng chính về số lượng sinh viên tham gia trên nền tảng Khoa học dữ liệu 365 là gì?
Define the key questions (Part 2)
In the previous lesson, we posed questions about student engagement that we'd like to answer with our
analysis. In the next few minutes, we'll formulate a couple more questions relevant to the topic and, like
before, give expectations on the outcome. Let's start with the following. Which are the most watched
and most enjoyed courses on the platform. Several things come into play when trying to answer this
question to its full extent. The most apparent approach would be to study the total number of minutes
watched from a course. This would tell us which courses and topics students are most interested in and
spend most of their time on. Such a metric, however, favors longer courses in terms of duration and
courses that have been on the platform for the longest time. Moreover, introductory courses at the top
of the course offering are expected to be among the topmost ones because of their position and
beginner friendly content. They're the most obvious choice to start from when embarking on a data
science journey. It would therefore not be sufficient to study only the total minutes watched. Another
metric that could be studied alongside the previous one is the average number of minutes allocated to a
student. Following the definition, this metric would increase with the number of minutes watched and
decrease with the number of students increasing. Its pitfall is that it favors longer courses and those on
the platform only for a while. The reason is that there wouldn't be many students who would have
started it, and a smaller number of students increases the value of this metric. Attempting to correct
these two metrics drawbacks, we can define a third one. The completion rate. Such a number would tell
us the fraction of a course students complete on average. How would we calculate that? Let's look at
the metric we've just discussed. Minutes watched per student. What could be the highest value of this
number for a specific course? Well, the length of the course itself, or in the case of students rewatching
the course, it would be a number greater than the length of the course. Therefore, dividing the minutes
watched per student by the course length, we'd receive a metric corresponding to the completion rate.
But unfortunately, this doesn't come without any biases. This metric would uplift shorter courses
instead of longer ones because shorter courses are much easier to digest. Additionally, the content that
is unlocked for free users is between 20 and 30 minutes. Which means that the fraction of unlocked
content to the length of the course is much more extensive for shorter courses. For example, a 30
minute unlocked content for a six hour course corresponds to about 8% of its length, but the same
amount of unlocked content for a course that is an hour and a half long corresponds to about 30% of the
course, which is considerably larger. This would result in a more significant completion rate for shorter
courses. No one accurate metric, however, would reveal the most successful course on the platform. We
should consider various factors when performing such an analysis. Whew. We're almost done. Only two
questions remaining. Another component of engagement is participating in exams. So let's ask the
following. What's the number of exams taken on the platform? What is student's general exam success
rate on the different types of exams, including practice course and career track? Watching courses on
the platform and solving the exams. Accompanying them is an integral part of the 365 product. It's
therefore essential to know what the level of engagement is for both components. We expect the
success rate of practice and course exams to be close to hypothesize about the performance on the
career track exams. However, let's first clarify what a career track is and how a student can pass it.
Career tracks represent a collection of eight carefully curated courses, forming a complete program on
either of three job titles. Data analyst, Business Analyst and data Scientist. To pass a career track. A
student first needs to have completed ten exams in total. Course exams of the seven compulsory
courses. Course exams of the two elective ones, as well as the final exam covering topics from all
mandatory courses in the track. The final exam consists of 42 questions and is evaluated similarly to a
course exam. 60% or more is marked as a pass. Upon completing a career track, students receive a
corresponding career track certificate, proving their expertise on various topics and opening new doors
to their desired job position. It's reasonable to assume that the number of people who attempt a career
track exam and the passing rate of these exams are lower than that of practice and course exams.
Career track exams are longer and cover a wide variety of subjects. And since we're on the topic of
exams, we can also study the certificates issuance. How many course and career track certificates are
issued? What fraction of the students who enroll in a career track complete it? The number of
certificates issued depends on the number of past course and career track exams. It's interesting to talk
about the career track exams as these are quite challenging to pass. Firstly, a student would need to
pass nine course exams to be allowed to take the final exam. Secondly, the track exam consists of 42
questions covering all compulsory courses. Therefore, it's expected that a tiny fraction of the enrolled
students would complete it. All right. We posed some interesting questions in this lecture and
hypothesized the possible outcomes. Now that we have our baseline in the form of postulated
questions, we can think about the Dashboard's appearance and the visualizations it would include. I'll
see you in the following lecture where we'll begin with the dashboard sketch.
Tại sao tỷ lệ hoàn thành có thể là thước đo sai lệch khi đánh giá sự thành công của khóa học?
Sketching the dashboard (Part 1)
We dedicated the two previous lessons to formulating the questions relevant to student engagement
content, watched exams passed and certificates issued. It's time to decide how to present this
information visually and arrange these visualizations in a dashboard best. We'll create a sketch of each
dashboard page. The first page would be dedicated to displaying a bird's eye view of the engagement
with the product. Let's start with three key performance indicators or KPIs. The number of engaged
students on the platform. The number of minutes watched per student. And the total number of
certificates issued. Let's reserve some space for the filters, including the status of the students, free or
paid, and the date on which we want to consider these KPIs. Next, let's position a horizontal bar chart
with each bar corresponding to a course and its length, representing the size of one of these three
metrics. Overall minutes watched minutes watched per student or completion rate. Let's also include a
filter showing the five leading courses concerning any of the metrics and the five courses that score last.
Such a restriction would be needed since offering all courses would look a bit overcrowded and the
chart will be challenging to read. Lastly, for this first page of the dashboard, let's include a donut chart
whose center shows the average rating of the platform. Its periphery would indicate the fraction of five
star ratings, the highest A course could receive four star ratings down to one star. Okay. I think that's
enough for the overview page. What's next? This one will be dedicated to the change of activity and
onboarding with time. Let's split the page horizontally in two. The top part would show the number of
engaged users versus time in a line chart where the user would have the option to filter between free
plan and paying students. The bottom part would represent the percentage of onboarded students
versus the registration date. Again, in the form of a line chart. Okay. One of the main advantages of
creating dashboards with software like Tableau is their versatility. We can create different views of the
data and give end users the ability to choose the period of interest and data granularity. With that in
mind, let's create several different views. The first would allow the dashboard user to choose the period
themselves.

The second view would split the periods into months, meaning users could choose the month they wish
to study. The x axis would then display the days of the specific month.
The third and final view would be monthly and the information would instead be delivered in the form
of a bar chart. Line charts are the go to when we need to plot many data points. While bar charts are the
preferred option when displaying fewer items. The reason is that in contrast to the days inside a month,
the 12 months within a year are not that many. So far. We've sketched the first two pages of the
dashboard and we have a couple more to go. In the next lesson, we'll discuss the visualizations they'll
contain.

Sketching the dashboard (Part 2)

In the last lesson, we sketched the first two pages of our dashboard. Here, we'll continue in the same
fashion, discussing a way of representing levels of student retention. One way this can be done is
through a cohort analysis table like the following. This visualization is not as straightforward to explain
as a bar or a line chart. A discussion we'll have later when constructing cohort tables in Tableau. I'll do
the same for the so called retention curves. A line chart closely connected to the cohort analysis. We'll
place these two charts on our dashboards. Third page, reserving space for a student type filter and a
color legend. Good.

Moving on to the fourth page of our dashboard, we'll study the exams, attempted and certificates
issued. It would be interesting to see how many of these attempted exams have been passed and how
many have not. One visualization we can realize is a horizontal bar chart with each bar representing the
month of a given year and its length showing the number of exams attempted that month. The left
section of each bar would tell us the percentage of exams that have not been passed while the right
section would show the opposite. The second half of the page would enable toggling between two
charts. The first would be a conventional vertical bar chart where each bar displays the number of
certificates issued by month. Let's allow for filtering based on the type of certificate, a course or career
track. The second would be a funnel type visualization realized as a horizontal bar chart. The top bar
would represent the number of people enrolled in a career track. Out of these people, we'll pick the
ones who have attempted at least one course exam from the track and those who have completed a
course exam. Next. We are interested in the fraction of students who have attempted a final exam.
Finally, we'll ask how many of them have passed the final exam and as a result, have earned a career
track certificate. Such a funnel can be filtered by the type of career track. We have just a bit more work
left to do. Let's design the final page by again splitting it horizontally in two. We want the left part to
visualize the minutes watched by students each month and the average minutes watched each month.
We can do this with the help of a combo chart. The bars will display the overall watched minutes by
students, while the line will visualize the average minutes watched. It would be beneficial to see the
difference in engagement between free plan students and paying ones so we can add a parameter to
filter out these two categories. Well done. Now let's turn our attention to the right hand side of the
page.

Here will incorporate two combo charts. Showing the conversion rate of students as well as their
subscription duration. The purpose of this pair of plots would be to study the behaviour of different
groups of students based on the amount of content they've watched on the platform. The visualizations
will be again toggled with the help of navigation buttons. And we're done with the sketching part. These
visualizations will help us answer the questions we formulated earlier in the course. In the next section,
we'll start retrieving data from the database using SQL.
Section 03: Retrieving relevant data from the database
Types of data access

Section 04: Introduction to Dashboards

Size of the dashboards
We are now entering the practical part of this course where we'll build the dashboard in Tableau.
Exciting, right? I've opened a tableau sheet here, creating a simple chart for illustrative purposes. You
can find the workbook in the resources below. But this won't be the workbook we'll use for the project
itself. Before I continue, I assume you're familiar with Tableau's basics. What a sheet is. The difference
between a dimension and a measure. The mechanics of creating a chart. Beginner level calculations and
how filters in Tableau work. I also surmise you have a public tableau account where your work will be
stored. If any of this sounds unfamiliar, don't hesitate to check out our introduction to Tableau Course
or the relevant sections of the complete data visualization course with Python, R, Tableau and Excel.
Ready? Good. First notice that apart from creating a sheet, there are two more options a dashboard and
a story. We'll work with dashboards in this course. So let's click on that. We've created a new tab
Dashboard one. You can change the name by double clicking on it. I'll go with dashboard demo one.
Let's examine the dashboard pane on the left hand side of the workbook. We'll explore them each one
by one. Looking at the top left corner, we see a device preview mode. This allows the creator to choose
the size that best fits their dashboard. Depending on the device it'll primarily use. We can choose
between several desktop options, tablets or even phones. Additionally, we can also select the
orientation. For our purposes, though, let's choose our own dimensions. This is done in the size section
below. Once again, Tableau allows us to choose between options, including fixed size, automatic and
range. If I press automatic, a message appears reading the dashboard will resize to fit any screen it's
displayed on. But as reassuring as this sounds, this won't often be the case. Resizing elements obscures
the dashboard and results in unpredictable behavior. So we'll choose the fixed size option. Let's fix it to a
generic desktop size of 1366 by 768 pixels. All right. Going down the dashboard pane. We can see all
sheets we've created and can add to the dashboard. This happens through a simple drag and dropping.
Right now we have only one sheet at our disposal. But be patient. We'll have a bunch very soon. At the
bottom we can see a section titled Objects. These are all elements that can be put on a Tableau
dashboard. In the next lesson, we'll review the ones used throughout the course.
Horizontal, Vertical, and Blank objects
In the last lesson, I walked you through the dashboard view in Tableau and we fixed the size of our
exemplary dashboard. Now we'll concentrate on the objects section, storing all objects that can be used
in a dashboard. Let's get started. First at the bottom of the objects section, we can choose between tiled
and floating layout. But what does that exactly mean? The tiled layout automatically fits the elements on
the dashboard. While you must specify the objects dimensions with the floating layout. Let's walk
through an example with the tiled approach to understand the difference between the two better. In
the top left part of the object section, we see objects called horizontal and vertical. Both are what in
Tableau we call containers. They are called containers because they can store other elements inside
them aligned horizontally or vertically. Let's experiment a bit with those. If I drag a sheet onto the
dashboard, it'll fill up the entire dashboard. Let's now try and drag a horizontal container onto the page.
We see that Tableau offers several options as to where we can position this container. Let's position it
below the sheet. Since this is a horizontal container, let's try and place several objects inside aligned
next to each other. What could those objects be? Well, we see that there is a blank object. Unlike a
container. However, it's a solid object and we cannot put anything inside it. Select the blank object and
drag it inside the container. So did anything happen? How can we tell? The object is inside the
container. Let's switch from the dashboard pane to layout. We can see many options that we'll explore
later in this course. For now, let's concentrate on the item hierarchy section. I'll expand all levels until I
find the blank object. And here it is. We can see that it's indeed inside a horizontal container. Okay,
What's next? Let's go back to the dashboard pane and drag another blank object next to the one we
placed before. At first you might have trouble targeting the right spot, but it'll come with practice. How
can we tell? We've got both blanks in the horizontal container. Well, go to the layout pane, navigate to
the blanks and ensure they're inside the container. To make matters easy on the eye, I'll color the blanks
by clicking on one of them, going to background and choosing a color I like. Let's color the second blank
object to. We can also put borders surrounding the objects or the container. Let's do the latter. Go to
the item hierarchy, select the horizontal container, go to border and choose one. Great. Notice the
white space between the objects and the container. We can see why this is by clicking on any of the
blank objects. See the outer padding option and outer padding is the spacing beyond the border and the
background color. The current value is four. We can change it to zero to see the difference without the
padding. Now let's also set the padding of the second object to zero. We can see no space between the
blanks and the containers. That's rarely a desirable effect. So let's hit Ctrl Z twice to return to non-zero
outer padding values. Okay, very good. Let's also experiment with the vertical containers. Grab one from
the objects section and place it somewhere. I'll go with the top right corner. Now, since this is a vertical
container. Let's grab two blank objects and put them on top of each other. Again, finding the right spot
to release the object might be tricky, but don't get discouraged. I'll go to background in the layout pane
and choose a color for my boxes. Again, go to the item hierarchy. Select the vertical container and put a
border around it. Notice how the borders of the two containers overlap. I don't like this look very much,
so I'll change the outer padding of both containers to a non-zero value. Say four. Now we have them
perfectly separated from each other. Great. Don't forget to press Ctrl s to save your work in Tableau
Public. We did quite a bit of work, didn't we? In the next lesson, we'll continue exploring the objects in
Tableau.

Text, Image and Navigation objects

In the previous lesson we introduced horizontal, vertical and blank objects. Now we'll experiment with
some other ones. By the end of this section, you'll be an expert in containers and objects in Tableau who
can effortlessly introduce them in reports to create unique visuals. With that in mind, let's see what the
text object does. Grab it and drag it somewhere on the dashboard. For example, in the middle of the
vertical container. Let's write dashboard. Demo one. We can change the letters, font size, appearance,
color and alignment. Let's make it look like a title by enlarging it and bolding the letters. Excellent. Now
we can place an image. Let's grab the corresponding object and put it at the top inside the vertical
container. In the pop up menu, we see two options for uploading an image to the dashboard. Inserting a
file with the image or providing a link. I'll do the former, which is also the default tableau option. After
you've selected your desired image, click. Okay. We now have branded our dashboard with a demo logo.
Right click on the object and you'll see many options appear. Find the ones that read fit, image and
center image. The former would fit the logo so that it fills up the space of the image object perfectly,
while the latter positions the image so that it sits at the center of the object. We can also change the
object size by manually adjusting it or by typing the specific dimensions under size in the layout pane. W
stands for width and H represents height again in pixel units. Good. Finally, we come to the navigation
object. This next part is a bit tricky, so please pay attention. Grab a vertical container and sandwich it
between the two blank objects. Now grab a navigation object and place it inside the vertical container.
Next, grab another navigation object and place it below the first one. Let's look at our item hierarchy.
We have a horizontal container that from left to right stores a blank object, a vertical object, and
another blank object. The vertical container in turn, stores two navigation buttons on top of each other.
Great. Let me now demonstrate one reason why working with containers is meaningful. Find the
horizontal container where the blank, vertical and blank objects are stored. Right click on the object and
choose distribute contents evenly. Nice. We can see that all objects in the container now have the same
width. A convenient feature to have. If you want particular objects of equal dimensions, just put them in
a container and distribute them evenly. All right. Navigation objects are used to navigate between
sheets, dashboards and stories. What do I mean by that? Let's right click on the dashboard demo one
tab and press duplicate. Let's now rename this second dashboard to Dashboard Demo two. Change the
text box to read the same and go back to demo one. Double click on the second navigation object. We
should now tell Tableau where to navigate once the button is clicked. Let's navigate to Dashboard Demo
two. We can choose the style of the button, text or image. Let's go with text for now. In that case, we'd
need to write a title. Dashboard Demo two will do. We can also give it some background color and a
border. Very good. Clicking the navigation button in edit mode won't navigate us to dashboard demo
two. While designing a dashboard rather than using it, you'd need to press the alt button on your
keyboard and then click on the navigation button. We've entered the demo two dashboard, but we now
have no way of returning to demo one. We'll therefore edit the upper navigation button accordingly.
Like before. Double click on the button. Navigate to demo one. Right dashboard Demo one as a title.
Choose a background and a border. And press. Okay. Now hold alt click the navigation button and we're
back to demo one. Great. We can now do a little trick on demo one click to configure the upper button.
Let it navigate to itself. This will allow us to choose its color rather than stay in gray. Before that, let's
write a corresponding title and do the rest of the design. This time I'll color the button yellow. This
would indicate that we are currently in Dashboard Demo one. Let's do the same for the lower button in
demo two. I'll just quickly make that analogous configuration. And there you have it. We can now switch
between two dashboards via these wonderful navigation buttons. Finally, let's use this incredible feature
of containers to distribute the objects evenly inside it. I think now is a great time to save our progress in
Tableau Public. We've covered a lot in this lesson, haven't we? In the next one, we'll discuss the item
hierarchy as well as tiled and floating layouts.

Item Hierarchy
So far, we've become familiar with how objects in Tableau dashboards behave. Now we'll step our
knowledge up a notch and learn how to arrange our objects in an ordered and tidy manner. The item
hierarchy we've used up to this point seemed rather messy. That can become tough to navigate,
especially when the dashboard becomes more complicated, containing several charts, filters, and other
elements like titles, logos and navigation buttons. That's why we'll try and recreate the dashboard we
constructed, but in a much more controlled fashion. Often dashboards are arranged vertically with the
title and logos at the top, then filters below, followed by one or several sheets and finally buttons.
Navigating to the next and previous pages. So let's have our first element be a vertical container. Next,
we'll have two horizontal containers on top of each other. The top would store a vertical container with
the chart next to it. The bottom container would contain a blank object, a vertical container, and
another blank object. Now let's return to our vertical container stored inside the top horizontal one.
Inside, we'll place the image, the blank object, the text element, and another blank object. Lastly, we'll
put the two navigation buttons inside the vertical container. And that's it. That would be our item
hierarchy. It looks much tidier, doesn't it? I'll wrap up this short lecture here. Next, we'll learn how to
build this item hierarchy in Tableau.

Tiled vs Floating Layout

We've already created a logical item hierarchy for our improvised dashboard. The next step is rebuilding
the dashboard so that this hierarchy is realized. Let's do it. Go to Tableau and create a new dashboard.
Let's call it dashboard item hierarchy. In the layout pane, go to size and adjust the settings to a fixed size
1366 by 768. Now go to the objects section and change the layout from tiled to floating. In this way,
we'll be the ones who control the location and dimensions of the items, not Tableau. In the last lesson,
we started with a vertical object. Remember? Now grab a vertical object and place it on the dashboard.
Notice how it doesn't fill up the entire space of the dashboard. That's because we are using the floating
layout and the object's position and size should be manually specified. Let's go to the layout pane and
change the x and y coordinates of the position to zero zero. This will place the top left corner of the
vertical object at the first pixel. Let me quickly comment on how the dashboards coordinate system is
designed. As you already know, the origin of the coordinate system is at the top left corner. Increasing
the x coordinate moves the object to the right, as one is used to from conventional coordinate systems.
Increasing the y coordinate, however, doesn't lift the object but moves it downwards. It might be a bit
confusing at first, but with time you'll realize that this is much more intuitive to work with. All right. Now
that we've got that out of the way, let's get our object back at position zero zero. Next, resize it so that it
fills the entire dashboard. Since the dimensions we initially set are 1366 by 768. Let's change the width
and the height accordingly. Okay. Very good. A quick look at the item. Hierarchy shows that as expected,
we have a single item, the vertical container. Next, we need to place two horizontal containers inside.
Let me now show you a quick trick. Instead of having the horizontal object floating like this, we can
make it fill the entire vertical container by dragging the object onto the dashboard and holding shift
before letting it go. Nice. Right. Let's do the same with the second horizontal object. Drag the object
onto the dashboard press shift and find the right place to drop your item. Let's double check to see if
both objects are inside the vertical container. And indeed, they are. You might wonder why I didn't apply
this same trick with the vertical object dragging it onto the dashboard, holding shift and letting it go.
Well, let me quickly demonstrate on a new dashboard. Look at what happens when I drag the vertical
object in this manner. Seemingly there isn't any change, but if I go to my item hierarchy, I can see that a
tiled object has appeared storing the vertical container. I wouldn't want the tiled object there as it's
quite unintuitive and difficult to work with. I can avoid having this object there by floating my vertical
object onto the dashboard and manually setting its position and size. All right. Now let's go back to our
dashboard. Inside the top container place a vertical object and the sheet next to it. Next inside the
vertical container. Place an image object. A blank object. A text object reading dashboard improved. And
another blank object. If positioning the bottom object seems tricky. I like placing my object in between
two other objects, then moving the bottom one to its correct position. All right. We have a bit more
work left to do. Going down to the second horizontal container, we place a blank object. A vertical
container. And another blank object. Let me give all blank objects some color so we can see them better.
Now inside the vertical container, place the two navigation buttons. Let's now use the magic of
containers to distribute contents evenly. We do this for the large vertical container, then for the top
horizontal one. The vertical inside it. The bottom, horizontal one and the vertical inside. Fantastic. All
that is left to do to obtain the same dashboard as before is the design part. Change the colors, put some
borders, fix the appearance of the image and the text, and edit the navigation buttons so that they
navigate between two dashboards. But you already know how to do all these things, so please do them
as homework to practice what you've learned. I hope you can appreciate how much easier it is to
navigate this item hierarchy rather than the messy version we previously experienced. What's even
better is that we can give all these objects names. For example, we can rename the big vertical container
to dashboard. Then the upper horizontal. One, two. Image, text and chart. The bottom horizontal one,
two navigation buttons and so on with the rest of the items. All right. Great job, everybody. You've come
a long way and are now prepared to embrace the task of creating a fully functional dashboard.

Section 05: Overview Page

Section 06: Engagement & Onboarding Page
Section 07: Engagement by Cohorts Page
Section 08: Exams & Certificates Page
Section 9: Content Consumption Page
Section 10: Discussion
Platform Engagement
Onboarding & Retention

Content Consumption

F2P CR and Subscription Duration

Courses Engagement

Exams

Certificates

Further Steps and Improvement