Data Visualization and Communication with Tableau – Week 1

Human’s have a propensity for patterns and stories. Through utilizing visualization tools such as Tableau it is possible to turn otherwise dense and cluttered data into a beautiful business narrative. These narratives can be used to recommend changes and improve business processes. In this course the student will learn how to effectively communicate business-relevant information mined by data analysis.

visualization timeline
Data visualization timeline

Ten Tips for Becoming a Data Analyst

  1. Ask questions, nourish curiosity, and embrace the unknown
  2. Start thinking about everything you see as a dependent variable or independent variable
    • “If I wanted to know more about this relationship, what would be the dependent variable, and what would be the independent variable?”
  3. Start exploring the advantages of Continuous vs. Discrete Variables
    • Continuous variable – Can be any value in between a minimum and maximum value
    • Discrete variable – Can only be a specific set of values (democrat vs republican)
  4. Listen AND Contribute – It is important to understand the business landscape so your recommendations are aligned with company values and intuitions
  5. Train your skepticism muscles
  6. SEEK details
  7. Cherish precision
  8. Best practices do not equal common practices
  9. Expectations matter! – What you present needs to be somewhat aligned with expectations or it will not be well received.
  10. Put yourself in other people’s shoes – Understand what your partners and stakeholders think when developing your solution.

Asking the Right Questions

A study asked business leaders to select which qualities they most desired in a new business analyst. The top three choices are below:

  1. Communication
  2. SQL and query skills
  3. Basic analysis skills

A separate study found that 99% of of data projects failed due to organizational issues and not technical restraints.

The aforementioned studies emphasize the importance of communication in the workplace. It is important to understand the context in which your stakeholders operate when generating your hypothesis and solutions.

Ask, Ask, Ask. Be curious and soak everything up like a sponge! Your job is to ask and answer as many questions as possible and understand the business landscape BEFORE beginning to analyse and interpret data, and certainly before you begin hunting for solutions.

SMART Objectives

In order for you be be successful you have to clearly define what business problem you are trying to solve. You need to understand the ins and outs of why the stakeholders are coming to you in the first place!

Before starting on a project you need to put yourself in the stakeholders shoes and ask questions to determine their true financial motivations. Unfortunately they may not always know.

Sample questions:

  1. What problem is this business having that you hope to solve by developing this project?
  2. Can you tell me more how this problem is affecting the business?
  3. What is your ideal outcome for this project?

Once you have your answers it is time to set your SMART objectives.

  • Specific – Identify what you will be doing as your analysis.
  • Measurable – How much would my business metric change if my recommendations are put in to action?
  • Attainable – Confirm systems are in place that make your goals realistic
  • Relevant – Make sure that the proposed project actually will impact the underlying business goal.
  • Time-Bound – Include a proposed time line for when the project will be completed.

Example: “In 2 months, analyze archived click-stream data to determine the website changes that will most efficiently increase revenues by 15% on a month-by-month basis compared to the same month last year.”

Listening to Stakeholders during Elicitation

The process of elicitation, drawing out or bring forth information from people, is an important step throughout the project life cycle.

During elicitation sessions the business analyst is attempting to achieve three main goals:

  1.   Identify your key stakeholders – Anybody who is affected by your data analysis project or who might have a strong interest in it.
    • Questions to ask yourself to ensure you have a succint list of stakeholders:
    • 1. Does the stakeholder have a fundamental impact on your organization’s performance? (Required response: yes.)
    • 2. Can you clearly identify what you want from the stakeholder? (Required response: yes.)
    • 3. Is the relationship dynamic — that is, do you want it to grow? (Required response: yes.)
    • 4. Can you exist without or easily replace the stakeholder? (Required response: no.)
    • 5. Has the stakeholder already been identified through another relationship? (Required response: no.)
    • Source: https://hbr.org/2014/03/five-questions-to-identify-key-stakeholders/
  2. Identify independent variables to test –
    • Question example: What do YOU think would help solve this problem?
    • Talk to everyone involved and determine what has or hasn’t worked already.
  3. Determine whether stakeholders agree about the problem to be solved  – your success is dependent on stakeholder buy in so you better make sure they are on board! The problem assigned to you may not be the actual problem the stakeholders experience.

To be a successful business analyst it is imperative that you keep tabs on which stakeholders participate in the context of your business problem. Keep asking yourself how you can turn what people say into variables you can test.

Stakeholder Expectations Matter

It is important to understand the company’s data culture before offering solutions. Business leaders can be reluctant to adopt more complex data strategies and instead opt for traditional methods or what the company has historically used.

Doug from Garner analytics breaks down the different types of data analysis into 4 types on his “analytics continuum”.

  1. Descriptive analytics – What happened?
  2. Diagnostic analytics – Why things happened?
  3. Predictive analytics – What is going to happen?
  4. Prescriptive analytics – Recommendations of what should happen?

Using SPAPs to Structure Your Thinking

At this point the business analyst should have already identified their dependent variable through SMART goals and their independent variables through elicitation sessions. However, before jumping into the data analysis it is important to set up an analysis plan. An analysis plan will help the analyst articulate what you are going to do, stay on track, get feedback along the way, and have a framework to understand the insights you find.

  1.  Add your SMART goal to the top of the pyramid.
    • Directly below your goal add the various dependent variables you are going to test.
  2. Add all independent variables (categories and subcategories of issues) you gathered during the elicitation sessions to help organize the various factors you will be exploring. There may be multiple layers will



Using SPAPs to Create Insights (SPAP part 2)

  1. List which data you will you use to address each sub-category. What data you will use from which data source. If you don’t know where you will get the data from, mark it on the SPAP!
    • This helps you identify if there is data missing or if there is data that will take you a long time to prepare
  2. Prioritize each category based on, impact, how feasible they will be to asses, and who suggested them
  3. Reorder the path based on the priority
  4. GET FEEDBACK from stakeholders. At this point it is a great idea to get feedback from stakeholders and make sure priorities are structured accurately. This is important to ensure you are using your time efficiently.
  5. GET DATA CRUNCHING: Review your data and look for large effects on your dependent variable for each of your independent variables. Once you identify which has the largest impact, dive deeper.
  6. Describe what your chart would look like, stick to line, bar, and scatter plot graphs.
  7. As you review the graphs look for patters and ask the following questions:
    • Do any effects stick out to the eye?
      • If not, mark off and make sure not to continue down the pyramid.
      • If yes, mark for later investigation
  8. Start of the top layers of your variables, if there is no effect at the top layer it is unlikely a lower level will have an effect.

Continue updating your pyramid as you gain insights and continually ask for feedback from all stakeholders! Once you identify which variables have the largest impact you can begin developing solutions.

This course will focus on how to analyze data in tableau to both develop and communicate insights. Stay tuned!

Leave a Reply

Your email address will not be published. Required fields are marked *