Move beyond traditional BI to empower your business faster and smarter than ever before. Data Analytics Unleashed brings you an ensemble of Alteryx, Power BI and Tableau Data Analytic All-Stars ready to share their stories and use-cases with you. If you're a data geek then this is the presentation you've been waiting for!
2. Ann Jackson
PHOENIX, AZ
@annujackson
Joseph Barth
PHOENIX, AZ
@sql_da
Michael Perillo
PHOENIX, AZ
@michaelperillo
Treyson Marks
PHOENIX, AZ
@treysonmarks
Data Analytics Unleashed
COMMUNITY USER GROUP LEADERS
3. Data Analytics Unleashed
FEATURED PRESENTERS
Brandt Wilson
PHOENIX, AZ
@brandtwilson15
Curtis Harris
SALT LAKE CITY, UT
@harris7curtis
Melissa Kovacs
PHOENIX, AZ
@firsteval
Erik Miller
DENVER, CO
@ermiller
Alex Duke
CINCINNATI, OH
@alexduke
Adebayo Adedeji
PHOENIX, AZ
@bayoshoe
4. Phoenix Tableau User Group
Join the Phoenix Tableau User Group every third Thursday of the month.
http://community.phxtug.com
http://linkedin.phxtug.com
http://twitter.com/phxtug
5. Phoenix Alteryx User Group
Join the Phoenix Alteryx User Group every third Tuesday of the month.
http://community.alteryxphx.com
http://linkedin.alteryxphx.com
http://twitter.com/alteryxphx
6. Phoenix Power BI User Group
Join the Phoenix Power BI User Group.
http://www.phxpug.com
http://linkedin.phxpug.com
http://twitter.com/phxpug
7. Arizona SQL Server User Group
Join the Arizona SQL Server User Group.
https://www.meetup.com/Arizona-SQL-Server-User-Group
10. Brandt Wilson
• Born and raised in Arizona
• Served a Mission for my church in
Bolivia for 2 years
• Married in December 2015
• Graduated BYU-I in July 2016 with a
degree in Business Management and
an emphasis in Supply Chain
• Been at Knight since January 2016
17. Director to VP to COO
• “There is no way the numbers are
this low”
18. Data Pushes the Needle
• One Week Later
• Our Pre-Trip Inspection Percentage rose over 20%
19. More and More Reports
Total Report Time: 2 – 3
Hours
• Yesterday’s Pre-Trip Percentage
• How many trucks drove/how many did a
Pre-Trip
• Trying to push past 70%
• Last 7 days who didn’t do a Pre-Trip
• Last 2 weeks who didn’t do a Pre-Trip
• Last 30 days who didn’t do a Pre-Trip
24. Build a workflow then just click play
• “Build it out kind of like you would in excel one time and from there
on out just plug the new data in.”
25. 1.5 Hours to 10 minutes
Total Report Time: 10-15
easy minutes (waiting for
website)
• “Stop thinking in excel and tables start thinking in workflows, one
thing leads to another.”
27. July 2016 – Connected to the Data
• No more waiting on a website.
• I could connect to our AS400/company data.
• Even more of time was freed up, to focus on other things.
Total Report Time: 5 easy
minutes or less (creating the
visual)
29. Power BI Pre-Trip Inspections
https://app.powerbi.com/groups/31898c7f-6be3-47b3-8a14-
a8cf6076e5ae/dashboards/9ffe75a3-e826-46e8-8404-
2daa0f42bc15
Total Report Time: 0 seconds
33. November 2016
• Saving Man Hours
• Trusting Technology
• The most dangerous
phrase in the language is
“we’ve always done it this
way” Rear Admiral Grace Hopper
34. PTI Alerts
• Alteryx allowed me to build out the process quickly
• Sped up the development process
• Every hour the alerts went out
35. Alert Results
• 2 days of alerts - 75% to 90%.
• As we were implementing and trouble shooting the alert system, IT
asked me to audit the process using Alteryx
42. Senior Business Intelligence Developer | Health
Catalyst
2016 Tableau Iron Viz Champion
2017 Tableau Public Ambassador
Twitter: @Harris7Curtis
Blog: www.themarkscard.com
Tableau:
https://public.tableau.com/profile/curtis.harris#!/
43.
44.
45.
46.
47.
48.
49. How do we make Tableauing easier for everyone?
70. Data for Good: Visualizing Data in the Social Sector
Data Analytics Unleashed
April 18, 2017 | Phoenix, AZ
Melissa Kovacs, Ph.D.
Founder, FirstEval, LLC
mkovacs@firsteval.com
@firsteval
71. What is the Social Sector?
(and what do they do with data?)
3rd Sector:
Social Economy
Private Sector
(Corporate)
Public Sector
(Government)
Nonprofit Organizations
Foundations
Volunteers
Philanthropy
Community
Social Mission-driven orgs.
Civil Society
72. What is the Social Sector?
(and what do they do with data?)
3rd Sector:
Social Economy
Arizona’s Nonprofits…
• Generate more than 8% of the state’s Gross State Product
• Are AZ’s 5th largest non-government employer
• Contributed $22.4 billion to the AZ economy in 2014
• Number 21,137 organizations
Source: Arizona Nonprofits: Economic Power, Positive Impact, Feb. 2016, AZ Community Foundation.
73. What do Arizona’s Nonprofits Do?
Source: Arizona Nonprofits: Economic Power, Positive Impact, Feb. 2016, AZ Community Foundation.
74. Data Is A Mess!
Social Sector Data Challenges Mirror the Other
Sectors’ Challenges …
76. http://www.hominc.com/hom-inc-dashboard.html
There is only one cure for homelessness. Just one. HOUSING. We like to think we're
pretty good at operating permanent housing programs for individuals and families who
are working with us and our partners to END their homelessness. Our core area of
expertise is the administration and management of tenant based rental assistance -
in Permanent Supportive Housing (PSH) and Rapid Rehousing (RRH) models - both
with a Housing First approach. We excel at getting people housed efficiently and
effectively, all while providing unparalleled customer service to our program
participants, landlords and service provider partners.
77. Thriving Together is an initiative designed to better prepare a quarter-
million young people within the greater Phoenix metropolitan area
for success from birth to career. The unique initiative brings together
resources from throughout the community to help students improve
educational results as they reach key benchmarks in their
development, placing and keeping them on a path to success.
Metrics: state test assessment scores, high school graduation rates,
…
83. The Setup
• Go to technical conference with DBAs, Tech
Architects, etc.
• Change settings on phone to match access point
• Sit in the middle of the room
• Wait…
• …but not for very long
• Thank you for falling for my trap
84. • People couldn’t tell what they were connecting to
• 77 people in the room
• 10 were pwned (13%)
• I didn’t take credentials
• I didn’t take data
• But it was TOO easy!
The Outcome
85. Why it matters
• People are the weakest security link in any company
• Many security threats are introduced by people
who are just not security conscious (ex: Target,
Home Depot, etc…)
• Once a malicious actor is in your systems, they may
not be discovered for days, weeks, or months
• Our job is to cut off paths through systems…
87. • Building program to measure risk across
Western Union
• Growing operational & big data analytic
concepts in the information security
world
• Run Western Union’s Governance, Risk &
Compliance system
• Manage three person cyber security
analytics team
• Manage Alteryx Server for WU
• Work with multiple groups to get their
analytic processes off the ground
• Bonus fact: I’m a Phoenix native!
Who am I? Sr. Systems Engineer – Cyber Security Analytics
88. • Telegram / Telegraph company
• Founded in 1851
• Financial Services
• 200+ countries
• 20,000+ employees and contractors
• 500,000+ global Agent Locations
• $150 billion moved globally in 2015
• Online & mobile presence
• 31 transactions per second
• Bonus facts: First telegram sent in 1851;
last telegram was sent in 2006
A little about Western Union
89. • Small team of three
• Chris Rodgers – Currently focused on
Agent Location analytics
• XinYuan Liu – Currently focused on
employee behavioral analytics
• Erik Miller – Currently focused on
Identity & Access Management analytics
• Analytics for most people in Information
Security doesn’t mean what you think it
does…
About the Cyber Security Analytics Team
Erik
Chris Xinyuan
90. • Our team started out as an army of one
• Approached to run a POC on Agent analytics
• Painful process with Excel, Access, MySQL, and 36 input files from six different
sources (Netezza, Business Objects, fraud analytics system, credit & risk
reporting, transaction pattern system, and global Agent listing)
• 100 hours per month spent on producing a risk profile of our 500,000+ Agent
Locations, 200+ countries, over 2 million POS Terminals, 13 security measures,
and all fraud in the past 365 days
• 4,000,000+ records each month
• Now we’re a full fledged Security Analytics shop
Our Start
92. • IAM means
– Identity (and)
– Access
– Management
• One of the most important, often downplayed
parts of Information Security
• IAM for WU consists of several parts, today,
however, we’ll cover four:
– Access Requests
– Entitlement lifespan (birth to death)
– Privileged Account Management
– System access monitoring
IAM? What Is it?
93. • Our first steps
– Initially started out as a discovery process
– “Just pull some data and see what you find”
– Realized that the systems had HUNDREDS of tables
– Realized the process owners didn’t understand their
systems
– Realized that the system vendors didn’t understand their
systems
– Realized that no one had really looked at the data in these
systems…ever.
– Realized that our report builders didn’t know the systems
– Realized this was going to be REALLY painful
– Realized that we could really, really, REALLY help!
IAM: It’s kinda painful and awkward
94. • Access Requests
– Who is requesting what?
– Are we confusing people who request things?
– Are requests being rubber stamped?
• Entitlement Lifecycle
– Who actually has access to what systems?
– Do they need it?
– Are we taking away access when we need to?
• Privileged Account Management
– Are highly privileged accounts being managed?
– What % of our servers are managed?
– In which domains?
– Where are we falling out of compliance?
We knew what we wanted to do...
95. • To find ties between all of our various tables – in
PAM, in request management, in identity
warehouse
• To determine which columns we could use to tell
stories…
• …and rule out those that are noise
• We created scheduled reports with Server to
show what we’re doing well…
• …and where we can improve
• And, finally, we use it to document our journey
through the IAM team’s data and our
recommendations
• And once that was done…
So we called on Alteryx first…
96. • To give us a visual presentation layer to better
highlight peer access
• To show when access makes sense…
• …and when it is questionable
• To tell the risk story of our critical system users
• We didn’t aim for “pretty” charts, we aimed for
immediately actionable intel
• Not every chart has to be a work of art…
• …but it should have a story to tell and have a call
to action
We then called on Tableau…
97. • For easy & inexpensive report distribution
• For metrics for the management of systems
• To tell stories about our user base
• We allow people to create their own
dashboards….
• …and their own reports and discoveries
• We make sure the data being used is the RIGHT
data
And we round it all out with Power BI
98. • The systems/processes we have been working with in the
last couple of months have been around for years and
years…
• …and years
• Until now, the groups managing the systems didn’t know
what power they had in their data
• In the last few weeks, we have:
– Been able to identify all users who have transitioned roles and not been
relived of their previous entitlement
– Found 4,200 entitlements not used, reducing the confusion of system
owners, end users, and our team
– Identified 1.7 Million accounts on servers at WU, which were managed
by Privileged Access, which were not, and what they typically do
– We were able to document, analyze, and metric the successes &
opportunities of our IAM program
– We were able to create a data dictionary on all of the systems –
something which even the vendors didn’t have!
Alteryx + Tableau + Power BI
99. • Additional employee / contractor behavior analytics (in progress)
• New development on our Agent Network analytics (in progress)
• Employee job, geographic, title end point usage analytics (Starts Monday)
• Server & asset cataloging and anomaly detection (Starts tomorrow)
• Consolidated risk metrics for all programs (in progress)
• Machine Learning on our massive data sets (in progress)
Our future initiatives - 2017
101. The Analytics Talk Podcast
• A podcast for analysts
• By analysts
• Panels of amazing speakers…YOU!
• Multiple viewpoints, one topic
• Beginning in May
• Come see me anytime today for a contact
card
• We want YOU!
Visit us, answer the survey, get involved!
Blog: www.theanalyticstalk.com
104. It’s My Data and I Need It
NOW
Self Service Analytics at Road ID
Alex Duke
@alexduke
105.
106.
107. 1
Special Sauce
Our customer service will blow your socks off. We know that EVERY customer
interaction is important. You'll find that Road ID is not just another faceless company. We
are real, down-to-earth people with personality, heart and soul. We will treat YOU the
same way WE want to be treated. It's just that simple.
108.
109.
110. 2
Do it Right
No short cuts. No excuses. We stand behind everything we do and every product we sell. If we
make a mistake, we make it right. Our reputation is our bottom line.
113. 3
Ear to the Ground
Eye to the Sky
Listen, identify, improve and innovate. We realize that most great ideas come from those closest
to an issue. We listen more than we talk so we can identify potential problems or areas that
need improvement. We seek out problems and help solve them. We drive evolution in
everything we do.
114.
115.
116. 4
Share the Love
Giving Back is not a secondary obligation. Instead, it is a fundamental reason to be in business.
As an example, "Road ID Gives Back" is an ongoing program where we pledge to donate a
portion of EVERY order to one of 12 excellent causes. As a fun twist, we let YOU tell us which
charity should benefit from YOUR purchase.
118. 5
Get on the Bus
We are a great company because we have a team of outstanding individuals that work
hard to achieve common goals. We strive to treat each other as well as we treat our
customers. We have faith in what we do and are passionate in our execution. We
encourage and celebrate hard work, creativity, productivity, having fun, quality work,
efficiency, strong attention to detail and leadership by example. Having great people
enables us to empower our teams to act entrepreneurially, make decisions and to take
informed, responsible risks.
119. Web Authoring
Web Authoring allows users to create dashboards within Tableau Server –
no Tableau Desktop necessary –
either from scratch or within a workbook.
121. Forecasting
Use Data Mart to
quickly grab inventory
data
Forecast for the
Group
Determine Current
Product Distribution
Apply Distribution to
Forecast
124. 6
Whoop it Up
Yeah, Road ID is a serious product and we have some very serious goals. That doesn’t
mean we have to be boring or stuffy. We have fun, we’re a little weird and we laugh as
much as possible. We realize that if we take ourselves too seriously, nobody else will.
133. 133
I am not
- A Programmer
- the king of Coding
- an office Night Crawler
- Not a computer Whiz
Bayo Adedeji?
I am
- A normal finance guy
- Business solution driven
- Work Smart guy
- Lover of time on couch
135. 135
What I do
Assortment planning and Optimization
Product
Information
Fixture
information
Brand
Information
Marketing
Information
Linear
Footage
Store
Count
136. 136
What I want to know?
How am I
doing? $$
How high is
high? $$
What does
the customer
think? $$
What sells
with what?
$$
Can I? $$ Should I? $$
137. 137
How I analyzed before
Planogram
Database
Floor Planning
Database
Product
Database
External
Data Sources POS
Database
Enrich
Using multiple UI and products, we download data into files and manipulate
manually
139. Deeper Understanding
• Industry standard behavioral clustering to drive incremental margin
• Provide customer insights, demographic information and location based
analysis
Speed and Accuracy
• Complex Efficient
• Early visibility to decision impact
• Fast error checking
Scale and Flexibility
• Analyze large amounts of data to drive sales
• One store at a time in-depth analysis
• Merge data sources and clarify actions
Variety of Usage
• Multipurpose tool with versatility
• Anything related to synthesizing data
139
Immediate Return
140. 140
Alteryx vs. Legacy Approach
Performance
Distribution
by Location
6
hours
10
Mins
Demand vs
Supply
(800M rows
of data)
2.5
days
55
mins
Bill of
Material
Estimate
(134k fixture
compare)
3.2
days
15
mins
LCR
Process
(21M Store
Sku
compare)
7.1
days
15
mins
Behavioral
Clustering
with
Constraints
6
Weeks
45
mins
Manual
Alteryx
Process
143. 143
Grouping Stores together based on behavior around a unique product attribution
Statistical Analysis (Clustering)
144. 144
This is an example of output of a behavioral store cluster with no constraint
Analysis Behavioral Clustering for Localization
145. Location Analysis (Spatial Analysis)
Spatial Analysis for specialty pet understanding the location of top stores as
it aligns to migratory patterns
146. Location Analysis (Spatial Analysis)
Top 100 Chicken
Stores
Spatial Analysis for Chicken understanding the location of top stores,
distance between stores that have chicken and potential ways of
maximizing locations where it is successful
All Chicken Stores
147. Alteryx + Power BI
o Consolidation of separately managed Excel Files
o 22,000 Formulas Pointing to 26 files on the network
149. Alteryx + Power BI
o Two Separate Reports that didn’t tie
(late nights)
o 4 hours per month on Excel
consolidation and report generation
o W/Alteryx, it now takes 1 minute
156. 156
Project Accomplished
Business
Insights
Behavioral Analysis
Space optimization
Location based analysis
Predictive Modeling
Data
Comparison
Product Flow analytics
Promotional Space
Analytics
Pre Bill of Material
estimations
Early Inventory Liability
Estimation
Analytics
Localization and Clustering
Predictive Demand
Modeling
Customer Decision Tree and
Dendrogram generation
Pricing Collaboration
Data Blending
The two main purposes of my presentation are to One, show the power and effectiveness of data analytics as well as these tools. And two, show the opportunities that these tools can provide for the user.
In order to clearly explain the power and effectiveness of these tools, as well as the career opportunities they can provide I will add things to this time line throughout the presentation.
For example, as I mentioned earlier I started at Knight in January 2016
-Only thing I had as an intern were excel skills and an almost finished supply chain/business degree
-Hired Originally as an Operations/Planning Intern
-Hiring Freeze, last minute switched to be a MobileComm/Safety Intern
- So I really Came in knowing nothing of what I would be doing
-No access, no training, no problem
-Computer/Excel
-Lots of data 4 spread sheets
- You have experience in excel right, lets make a report to know what is going on with Pre-Trip Inspections
A mandatory vehicle inspection that each driver must do before each Trip. He/She checks out their truck and trailer to make sure it is safe and ready to drive.
Zonar, RFID tags (how long it takes, time from tag tag, when the inspection was done, etc)
This is day 3 or 4 of my internship, I am now in the full swing of getting trained answering phone calls etc, during down time and late afternoons had time to work
**show screen shots of GTC website**
** 4 reports in a csv files, daily**
***Show Report***
-When I showed my director the numbers his jaw dropped
-He brought me into the VP’s office
-And he called the COO in.
-I was assigned getting these Pre-trips up as my main focus project
-Within less than two weeks the intern had met the VP and COO and
-Now reported to both of them with weekly numbers/progress
-Why are Pre-Trips so important other than just safety…. The industry leading causes in CSA violations (tires, lights, and breaks) (all preventable with a quality Pre-Trip)
explain why this is such a big deal, and men
1/15/2016 - 2/1/2016 (+20%)
After sending the report out each morning for one week
????DMs, DDMs, operations etc????
Showing the right data to the right people is key
Trying to push past 70%
Some of these reports were important to certain people but not everyone and not constantly.
If anything they were probably taking away from each other.
The data helps us discover an issue we did not fully understand
Driver doing Pre-Trip correctly not getting credit
***show failed inspection report***
This Data was no included anywhere
Dirty/Messy Data (no asset number)
Starting stopping etc ( answering phone calls etc)
Started Tracking Failed Inspections and adding in numbers to Pre-Trips
2/1/2016 – 4/15/2016 (+5%)
At this point you could have a large portion of each day was in excel
From teaching people how to use filters, vlookups and pivot tablets.
To trying to automate and improve existing reports.
I was preparing to go back to Idaho to finish my last trimester of college.
By the end of my internship we had gained another 5% on most days, but fluctuated quite a bit and dropped off a lot over the weekend.
Offered a Full Time Position( to work from home, until I graduated)
Born and raised in Arizona, great university never could get used to the winters, end of fall, start of spring.
Once there due to our lack in staffing at the time in the mobilecomm department I was not always able to get daily reports out.
The PTI % began to decrease
The percentage varied depending on the focus from leadership as well as the reports I was able to send out that day, on any give day we were down 5-10% from
6/1/2016
It was probably around June 1st, 2016
By this point I was very frustrated and discouraged I hardly got anytime to do anything but answer the MobileComm phones
I felt that I had taken a huge step back in pushing the needle and making a difference
Over the phone our New Director of Data Analytics Treyson Marks, and some people from Alteryx introduced me to the program.
Alteryx was still very new for our company and I had heard nothing about it
While giving me a demo of what Alteryx can do a couple things really jumped out to me.
This first thing that really grabbed my attention was the power of this little green button.
???Build Graph of time??? In excel???? Power bi???
Easy minutes = checking email etc
After completing my first Alteryx Workflow a report that took me about 1.5 hours took me as long as it took to download the reports from Zonar and plug the results into the excel visual
+1.5 Seconds give or take that alteryx took to complete the report
All my reports followed a similar process
This was all without even being able to connect to data
After I got back from Idaho things began to change rapidly
We added more people to MobileComm
And I began getting much more into Alteryx
I started learning basic SQL things and was given access to connect to our AS400 data
dashboards
This is as far in-depth as you could go
Full swing of things
Now I was connected to the data
Blending and organizing the data into usable tables
Producing high quality visual reports/dashboards.
When our Senior Vice President of Safety realized how much more we could do with this I was quickly asked to convert all safety reports.
(work in progress)
Received a promotion to be the MobileComm Team lead and Business Analyst for the Safety Department
KPI’s
Letting Rafael do his job
Etc
EXAMPLES OF PREVIOUS REPORTS
Others we didn’t have some these reports and visibility
Things I had done with data and the analysis showed leadership that I had some kind of an understanding of technical things
Leading Technology based Projects (used to fully lean IT)
Business Link to IT
One of the first big projects that made sense was PTI Alerts
Built the process out in Alteryx
If a truck moved, and did not have a PTI done, add a count
Based off of the Alteryx workflow and output I worked with IT and we developed an automated alert system, based on a count send a particular message to a Driver, and as the count grows send a message to the driver and a particular person or group of people, and keep escalating as the count grows.
My load has become extremely full with projects and reports, that paired with the company starting to embrace Data and these tools so much more, allowed me to step out of all other previous roles and be the first full time business analyst in the safety department
Anyone with the right desire and work ethic can do it
You might know me for this…
…but I’m here to talk about how we do Tableau at Health Catalyst
How many in the room have a corporate style guide?
Happy to be here, what I do, I spend time in the social sector space and in the for-profit world, and want to share my experiences with you today and walk through 3 use cases of Tableau in our local social sector. Contact me with questions.
Why we should care about the social sector? Huge part of our national and AZ economy, and a large employer. Mostly, though, we should care for the services they provide – nonprofits provide healthcare, education, housing, arts and culture, and numerous other services to all the humans around us. So how they do their jobs, and how they change their approaches to do their jobs better by using data or anecdote, is something we should care about.
Nonprofits use of data is largely evaluation, sometimes paid for by funders. Using it as a BI tool is just now becoming more prevalent.
Let’s dive a little more into what orgs in the social sector are doing with data…
Next, I’ll walk through 3 local use cases of Tableau in the social sector.
RBBB – collect lots of great data – in Excel, I worked to clean and analyze data and listen to their reporting needs, to get to these dashboards. This was part of a larger evaluation of the effectiveness of their tutoring program.
Used for two constituencies – internally and for clients
Self taught Mike
Comprehensive data center being developed in Tableau and will be released in late summer / early fall.
Also want to mention that it was tough to find these 3 local examples of orgs in the social sector using Tb. I may have missed some fabulous example, and would love to hear about more orgs using Tb. I can tell you from my experience many orgs are exploring Tb, but getting to the point of having influential publishable dashboards is harder. I don’t mean to say this to criticize the social sector, but rather to raise awareness of how difficult it is for nonprofits to get to using dashboards to influence their business decisions. Some of this is related to data access and data governance, but some of it is not having the analytic skills in house, or not having the resources to invest in analytics. Sometime scarce resources mean these orgs live off of excel spreadsheets.
Nonprofits don’t exist to make money. Their missions aren’t profit-driven. (stating the obvious). Vu Le is one of my favorite nonprofit thinkers and I think captures a spirit in this sector – working hard for social justice and transformation, and cooperation, all while living in a realm of scarce resources. This middle tweet I screen shotted here displays this attitude – nonprofit funders, like foundations, typically have a reputation for not wanting to pay for overhead, but collecting and gathering data is overhead, for an NPO – sometimes leaving data to being unfunded. Vu also taught me that there is science behind putting baby animal pictures into one’s presentations.