How to Turn Boring Visualization into Fascinating Data Storytelling.

6 min readNov 9, 2022

An actual case analysis illustrating how using Tableau Stories can help you convey your idea and keep your reader from drifting off or being confused.

Have you ever stared at a data chart and wondered, “What in the world are you trying to tell me with all of this?” Which is then followed shortly by another question: “What even is that intended to reflect?”

In this article, I’ll showcase how to use Tableau stories to transform a quite dull visualization into an engaging data story.

· Gather Data
· Explore and Assess the Data
· Data Features
· Data Limitations and Biases
∘ Data Collection
∘ Data Processing
∘ Data Insights
· Define the Problem Statement
· Improved Data Visualization

Gather Data

I chose a fascinating dataset from 2022/W42: The FDA Inspections dataset with the original data and visualization coming from the U.S. It was presented at MakeoverMonday, A research and data organization aimed to enhance how we view and understand data. Below is a screenshot of the shared visualization.

FDA Food Inspections Original Visualization

So, What is the problem here? Will…

They did not clarify what countries/areas are considered domestic.
The dashboard did not include context text like the data sources or the creator.
The dashboard’s text did not aid the end user’s comprehension (what do NAI, OAI, and VAI stand for, and what dose they mean).
The reader is not able to interactively explore the data within the dashboard.
The visualization was not centred on a specific, clear finding in the data.
The visualization did not include interaction or animation, The inclusion of filters and additional variables was not shown in the tooltip.
Colours were not used appropriately, Color choices must accurately reflect the data and be chosen with accessibility in mind. For example, values ranging from negative to positive numbers should be encoded with a diverging palette. Also, the colour palettes should work for colour blindness.

With all of these problems in the dashboard, but still, I must give credit where credit’s due, they used plots appropriate for the data types as they used Line plots for sequences, bar charts for categorical variables, etc.

Explore and Assess the Data

Go ahead and download the file “FDA Inspections.csv” from the link Here. The data set has 13 fields described below with 264640 recorded 217 from 2009 through 2022. This dataset has been compiled from data sets originally sourced from the FDA Inspections data bank.

Data Features

fei_number [integer]
legal_name [string]
City [string]
state [string]
zip [integer]
country_area [string]
fiscal_year [year]
inspection_id [integer]
posted_citations [string]
inspection_end_date [date]
classification [string]
project_area [string]
product_type [string]

Data Limitations and Biases

Whether you like it or not, the data you use for your work is flawed. It is frequently never a complete representation of the demographic you are trying to draw conclusions from. When making recommendations, it’s crucial to point out any potential bias points throughout the data collecting, processing, and analysis stages.

Data Collection

missing variables bias is possible because some records are missing state and zip code features, which is happening in the majority of countries other than the United States.

Data Processing

The list of variables that included outliers is presented below and you can explore them further in the following tableau story. but they have no discernible influence on the analysis’s findings.

Data Insights

Confounding Variables bias is according in this data set as the feature of Country/Area is set to the United States both State and zip are shown to be null, However, I discovered that the presence of this bias was minor. When missing variables are a bias introduced during the data collecting phase, confounding factors have an impact on the analysis.

Define the Problem Statement

Define what questions this dashboard will be able to answer.

The main reason for creating this dashboard is for exploring, the reader could be someone working at the FDA institution wanting to look at the bigger image of the institution or any non-FDA worker that wants to understand how the FDA inspections field is looking over the years in the US.

The audience of the dashboard can navigate the filters and find the answers to the following questions:

What is the overall number of inspections that have been done in the US in 2020?
Which classification out-turn from the inspections in 2020 had the smallest number of happing?
Which state has the largest number of Official Action Indicated classifications in 2020? How many inspections give OAI classification in that state?
What is the most inspected product type in that state in 2020? Did it change significantly from the previous year?
How many inspections in that product type on that state that got Official Action Indicated classifications in 2020 got posted citations compared to many that did not get posted citations?

Improved Data Visualization

Now that our questions are clear we can start drafting the dashboard wireframe, Although wireframes are frequently used in web development, I believe they are also quite beneficial in dashboard design. By keeping the concepts brief and always focused on the objectives of your target audience, you can work with wireframes much more flexibly than you would with anything designed in your dashboarding application. Below is the dashboard wireframe.

Wireframe of dashboard — The Dashboard Wireframe

Carrying out what the original dashboard did with their choice of plots appropriate and using line plots for sequences, bar charts for categorical variables, and maps for geographical data. but changing the colour palette to a colour-blind palette. Below is the full dashboard and you can explore it at the following link.

Now if you look at this gif of how the viewer can explore the dashboard and find the answers to the previous questions.

I appreciate your time. It’s been a lot of fun to share my thoughts and the skills I’ve gained via my data visualization journey. If you want to learn more about data visualization and how to begin with data storytelling, be sure to follow me on Medium and connect with me on LinkedIn.

Make sure to visit MakeoverMonday in the future to start your own data visualization journey because there will be a ton more updated data visualizations there.

if you want help with creating a dashboard in tableau contact me here https://www.fiverr.com/share/30VbYr

<! — Put this code anywhere in the body of your page where you want the badge to show up. →

<div itemscope itemtype=’http://schema.org/Person' class=’fiverr-seller-widget’ style=’display: inline-block;’>
<a itemprop=’url’ href=https://www.fiverr.com/alaa_data rel=”nofollow” target=”_blank” style=’display: inline-block;’>
<div class=’fiverr-seller-content’ id=’fiverr-seller-widget-content-78735b32–4249–484c-9c93–1e6af094c6bd’ itemprop=’contentURL’ style=’display: none;’></div>
<div id=’fiverr-widget-seller-data’ style=’display: none;’>
<div itemprop=’name’ >alaa_data</div>
<div itemscope itemtype=’http://schema.org/Organization’><span itemprop=’name’>Fiverr</span></div>
<div itemprop=’jobtitle’>Seller</div>
<div itemprop=’description’>Hi there, my name is Alaa and I am your one-stop destination for any data visualization or animation you may need. I have worked in information technology and computer science for over a decade. I am a data analyst who loves visualizing data and creating stunning, dynamic dashboards.

I am very flexible and open to a wide range of projects, so please don’t hesitate to contact me to discuss your project’s needs further. I have strong written and communication skills in English and Arabic.

[2] Monday M. 2022/W42: FDA Food Inspections. Data.world. Published October 11, 2022. Accessed November 6, 2022. https://data.world/makeovermonday/2022w42 ‌