本课程是《数据仓库和商业智能系列课程》的第四门课，建议按照顺序学习这些课程。通过本课程，你将学习到作为一个BI工程师应该具备的知识和技能，例如使用数据仓库实现商业智能的目标，你将有机会在数据仓库环境下使用大规模数据，还会使用MicroStrategy的在线分析(OLAP)和数据可视化功能来创建可视化的管理面板。 本课程概述了商业智能技术是如何支持在不同商业领域的决策制定，这些技术已经在公司战略、绩效管理和竞争意识等方面产生了巨大的影响，包括了广泛采用的决策支持系统、商业智能系统和可视化分析技术。课程按照商业智能概念、工具和应用的方式进行组织，包括使用数据仓库进行报表展示和在线分析，创建可视化和管理面板，以及业绩表现管理和描述性分析。 翻译: Shiny Zhu | 审阅: Yanhui Song Coursera Global Translator Community
Comprehensive Training Video for Creating Dashboard Using Advanced Visualization
Loading...
來自 University of Colorado System 的課程
商业智能的概念、工具与应用
350 個評分
University of Colorado System
350 個評分
課程 4（共 5 門，Specialization Data Warehousing for Business Intelligence）
從本節課中
Data Visualization and Dashboard Design
This module continues on the top job responsibilities of BI analysts by focusing on creating data visualizations and dashboards. You will first learn the importance of data visualization and different types of data that can be visually represented. You will then learn about the types of basic and composite charts. This will help you to determine which visualization is most effective to display data for a given data set, and to identify best practices for designing data visualizations. In lesson 3, you will learn the common characteristics of dashboard, the types of dashboards, and the list attributes of metrics usually included in dashboards. Finally in lesson 4, you will learn the guidelines for designing dashboard and the common pitfalls of dashboard design. You will need to use MicroStrategy Desktop Visual Insight to design a dashboard for a Financial Services company in a peer evaluated exercise.
與講師見面
Jahangir KarimiProfessor
Information Systems University of Colorado Denver
Hi, my name is Alex Jorna, I'm a Technical Account Manager with MicroStrategy.
And I'll be taking you through today's presentation, so let's start.
MicroStrategy is a business intelligence reporting tool that sits on top of your
data sources and provides a user with visual insight to your data.
Your data might be an Excel file, text file, relational database,
web sources, no SQL databases or
one of many other supported data sources that might MicroStrategy can access.
This is a product demonstration of MicroStrategy desktop,
which is a new interface that allows business users to connect to data
regardless of the type of source and conduct visual data discovery
on their own in a self-service way without the support of IT.
I'm going to start this demo by launching a sales dashboard,
a MicroStrategy file that has been saved to my laptop.
By launching this too,
you can see how it invokes the thick line that has been downloaded and installed.
This is a HTML5 interface and you can see how we have positioned the data
on the left and you can select all the visualizations from the right.
What you are seeing here is a dashboard that contains revenue information
plotted on a geospatial map, integrated with Esri, which provides out of the box
functionality and mapping capability with complete interactivity as you can see.
A user can readily drag and drop data onto a map to visualize
information like revenue, rather easily and quickly as well.
If you like to interact with data on visualizations that is not out of the box.
In this presentation, you can see there are several additional d
three visualizations that have been deployed,
which will also be covered during the exercises.
A user can add any third-party open source or
open source visualizations into your analysis and drag and
drop data into exactly the same manner and drill down on values.
So if you have revenue data and you want to visualize revenue over time,
you can do that as well and readily interact with the data to make it much
more an enriched experience and make some insights on the data that you have.
In addition to being able to add multiple visualizations,
one can also create interactivity between each.
As you can see, how I can highlight certain data points on the map and
use that as a filter with others.
So there is extreme interactivity between the visualizations that give
you that rich and analytical experience for the MicroStrategy desktop interface.
To add some additional analysis on top of that,
you can even bring in new visualizations.
So let's bring in one of the out of the box visualizations to see
revenue over time.
I can do that by drag and
drop of the relevant information from the dataset panel to the editor panel.
And instantly, I can plot all my revenue information by date.
You can see how quickly I have that information and I can observe any trend.
If I want to make the trend a little bit more evident and do it over a month, but
the data source does not contain the month attribute, I can enrich the dataset
by selecting the date and create a new time attribute by selecting the month.
With MicroStrategy, you can build these attributes called the reporting objects
on the fly, which really gives the analyst a lot more power to be self-serving.
By replacing the day with the new month, I now get the trend I am looking for
and more easily see the patterns within the data.
MicroStrategy provides a lot of functionality to the workflow and
exposes this through easy to use menu options.
For example, if I want to add a trend line, I can enable that
on revenue with a single click off of the right mouse click Menu on Revenue.
To further enhance the analysis experience,
we provide the ability to format the visualization.
For trend lines, there is also the option to do forecasting
by enabling it within the properties for the trend line.
By increasing my forecasting period, you can now see how I can view my forecast
beyond April 2015 data that I had and
make me as an analyst without training or
the experience to create complex analytical matrix more self sufficient.
This is a testament to how the desktop tool allows analysts to do some
enriched analytics on their own in a very self-sufficient manner.
This was done fairly quickly from a dashboard that already contains
some content.
To get started with the process for
creating a dashboard from scratch, we're going to through the steps.
So let's start by showing you, essentially what the desktop tool is to make
it easier with the following exercises that you will need to do on your own.
The process is fairly simple, you can connect to data source to bring the data
in and you can start building these realizations to gain inside to your data.
In case this is where you need to start with doing some data preparation,
you can do that as well making use of the data rangle feature
that comes with the MicroStrategy tool set.
This is the starting page,
you will notice I have some datasets already on the left-hand side.
To bring in new data, I will click on the add data option.
You will notice that there are several data source
options available within the desktop interface.
You can chose from any of the native options available or
we can choose from the database options connecting via OTBC connectors and
selecting from multiple database options and really bring in the data.
The key point I want to make here is that you have the ability to bring in data from
multiple different sources and the experience is exactly the same.
So whether you choose an Excel spreadsheet or a Haddop option, the experience
is the same and there is no leaning curve unless you can do that fairly quickly.
In this example, I'm going to start by choosing a cloud option.
I will choose mysalesforce.com account, it is a native connector.
Again, notice how I am not leaving the interface and
I am bringing in information from my Salesforce account.
It is one seamless workflow.
I login to my account and now I can browse my Salesforce reports,
drag and drop the data to bring into my analysis by bringing it in.
It readily converts information in Salesforce and brings in those tables and
columns and converts them into MicroStrategy console attributes and
matrix.
There is no data modeling involved and is a complete business user friendly tool.
Notice how everything has been done without any requirement for
data architecting or modeling.
Yet, it enables the business user to start going analytics on
their own In a very self serving manner.
Once you bring in the data, the next stage is to visualize the data.
But oftentimes it is required to prepare or cleanse the data, or
what we refer to as data wrangling within the toolset.
When looking at the data we can see we might want to make some changes,
like the Web-channel, and I want to remove that.
W want to click on the wrangle option which takes me to this graphical user
interface, where I have multiple options to tweak and
cleanse the way my data is looking, for improved analysis on top of it.
For instance, in my City column, there are blank fields,
and I can fill the data in, or I can standardize on the naming,
like Washington, DC versus D.C..
Here I can change that to look consistent, the way I want it, and
apply it to the entire table column.
Next we will remove the second part in my channel column.
As I highlight the section you can see the functions available in the top left
updates providing me with valid options to apply.
Anything I do is recorded as a history script, as you can see on
the top right section, which means that this is not a one time process.
But these steps can be applied the next time I bring in the data from sales force,
and I only need to create these steps the first time I am bringing it in,
making this a repeatable process for the future.
Let's do one more edit to my data.
I am looking at the Category column, and
will profile my data to see any spelling errors, or
if I want to merge or cluster the information to create consolidated values.
Again, this is done completely within this graphical interface, making
the changes on the fly and not breaking the process by taking the data into Excel.
These are just a sampling of what can be done within the wrangling feature.
There are dozens of functions and it's continuously being updated with
new features to the data wrangling, which has been integrated to work seamlessly
within the workflow process for business users to prepare the data.
Some other steps a business user can do within this workflow are to convert
the attribute and metrics from how MicroStrategy has automatically identified
them, for example, number of purchases, and days since last purchase.
All of these tasks are drag and drop based and easy to do in a very intuitive manner.
A key feature with MicroStrategy is that you can bring
disparate sources into a single in-memory queue for analysis.
Notice how I can connect to a different source and bring in information from
a Hadoop, relational data, or even an Excel file, allowing us to really start
making my analysis more complex, or as complex as I want it to be.
Yet simple enough to blend data and
increase the cross functional nature of my analysis.
Now you can see how I have tables from disparate sources in one single cube,
and I can create joints across them,
again, completely intuitive and very graphical in nature.
If there aren't any joints and I want to force them, like the first column of ID,
I would just drag and drop it onto the other column and
map it to instantly have the joined crater.
Now that I have defined my sources, made my connections and joints,
cleansed the data through the data wrangling option, you click on the finish.
And we'll be ready to start analyzing your data.
You'll be visualizing your data within the visualization editor.
Through building visual inside representations of your data
through which you can identify trends, outliers, or do forecasting.
In this case, the data that was brought in,
is some customer profiling information.
Let's say I want to build a bar matrix to see how my customer can be profiled
against age and city and identify how loyal they are by number of purchases.
As you can see, I can build a quick visualization on this matrix.
Let's say I want to enrich with colors to make it a bit more vivid.
I can do this by adding threshold colors.
Now you can see that Toronto seems to have more red than the other cities.
To improve the way I am seeing the observations, I can go ahead and
build a customer group on the fly.
Typically, this would require an administrator, to help with building this,
but with MicroStrategy the business user has the ability to do this on their own.
Select the items from the attribute and
create the custom groups in a seamless workflow.
This now brings the data point down to four, and
simplify the trend identification.
You might have other preferences.
Looking at this bar matrix, you might want to visualize it in a very specific format,
for which you identified this one visualization you have found on a D3
website, called a Sankey diagram.
Something as complex as this, and you would like to visualize your information
using this visualization specifically.
This might not be an out of the box visualization, but
it is existing open source visualizations that's available.
You can build plugins with MicroStrategy,
and during one of the future exercises, you will have the opportunity to
deploy a plugin to use within your visualizations of data.
To deploy a new plugin to be used, you will first save your work and
close MicroStrategy desktop.
For this demonstration,
I have a folder with plugins I have downloaded from GitHub.
I could have created them myself, as well.
I have the specific plugin I want to use.
So I'll just drag and drop that into the plugins folder
within my insulation folder, and bring it into my analysis.
Now that I have deployed the plugin, I will relaunch my dashboard.
And you can see that I have the option within the visualizations gallery, and
I can change from the bar matrix to this new visualization, and
view it exactly the way I want.
Similar to the out of the box visualizations, I can interact with this
new visualization in exactly the same way by adding dimensions and
increase the interactivity.
You can maintain that and use it as a filter if you will, and
enrich your analysis doing it exactly according to your preferences.
With MicroStrategy desktop, you can do some advanced analytics as well.
MicroStrategy provides over 300 analytical functions out of the box,
but similar to the D3, these functions can be extended and
enriched, to introduce R for statistical calculations and
algorithms to be applied to your visualizations.
For this demonstration, we will not go into the details of R, but
rather I would just show how this can be applied and make you aware of this option.
Let's start by adding a second visualization, as a scatterplot,
and create a loyalty program to segment my customers.
I can plug all my customers onto this visualization and
then create a model using R analytics to identify loyalty programs.
I can create this R model, based off of several different key attributes and
metrics, within my existing data source, and really build an advanced
analytical function, that is not existent within the out of the box functions.
We will apply this clustering algorithm model from R, which I will drag and
drop to apply it just like any other metric.
In a matter of seconds I have created groups or
loyalty classes to identify those customers that are loyal.
And I can create groups that I want to create marketing campaigns for and
allows me to do some cross functional analysis on top of my data.
In summary, MicroStrategy sets on top of your data sources and empowers
business people to analyze and understand data without specialized expertise.
With Analytics Desktop, it takes just a minute to create stunning,
interactive data visualizations and
data-driven stories to foster new insights and new understanding.
It provides the ability to analyze any data no matter where it is stored.
No complex scripting, modeling or SQL code is required.
