Visualizing India’s Public Health Infrastructure
Exploring India’s medical capacity using Data Humanism practices
Earlier this year, I was introduced to information visualization during my second semester at the National Institute of Design. The course was led by Prof. Chakradhar Saswade who guided us — his students — through our projects and connected us with professionals in the industry like Amit Kapoor and Prachi Nagpal.
During this course, COVID-19 had already entered India and there were questions about how it would affect the population and the health sector.
These questions included:
- Is India’s public health infrastructure equipped to handle an epidemic?
- Are the public hospitals in India capable of keeping up with the patient surges?
- Are there enough hospital beds in our state’s public hospitals in case of an emergency like a pandemic?
- Are there enough government doctors to attend to a rising number of patients?
India has majorly two types of health infrastructure: private and public. Private hospitals are expensive and poor households cannot afford them, which leaves public healthcare as their only available option. Initially, almost all the suspected cases of COVID-19 were reported to public hospitals and it was important to assess where the country stood in terms of medical capacity. Public hospitals are government-owned and play an important role in health care, providing care to patients who do not have access elsewhere. I set out to answer questions about India’s public health capacity with data visualization.
Data Gathering
My first step was to collect data from reports like NHP, Brookings, Wikipedia, WHO, The World Bank, etc., and collate them in Excel. I chose to limit my focus to the public health infrastructure part of the dataset. Now that I had my focus, I needed to clean and analyze the dataset. One of the biggest challenges in data visualization, that will consume most of your time, is cleaning the data. It includes taking big swaths of data and simplifying them to more basic and understandable terms. However, going too far with simplifying the data can lead to baseless conclusions or ignoring significant modifiers, leaving the reader with vague assumptions.
Planning the Narrative
While working on this project, I learned about information designer, Giorgia Lupi. I studied her visualizations from the book Dear Data, broke them down, and tried to imitate the forms and patterns she created. I experimented with the advice and suggestions she shared in interviews. Her manifesto, Data Humanism, inspired me to adapt an exploratory, data-driven approach to narrative development — an approach that encourages readers to freely discover when presented with information, as opposed to one whose goal is more rigid curation. This approach helped me build my visualization narrative.
Illustrating and Exploring the Forms
I started drawing to capture my ideas and plan the structure of information. I followed the three phases from one of Lupi’s interviews:
(1) Drawing to Structure
The first phase focuses on understanding the macro-categories to draw the layout or architecture of the visualization.
(2) Drawing to Explore Elements
In the second phase, you focus on the singular elements of the macro and define the shapes and colors that represent the data better.
(3) Drawing to Refine
The final phase can be called a draft phase, where you have decided on the components of the visualization and would like to refine it before going digital in the software to review and reflect.
I selected a shortlist of forms to test in Illustrator (see the illustration below). The form in [A] is based upon a seashell. Seashells are a symbol of protection, resurrection, and birth. However, it was difficult to read the data using that form. The form in [B] is based upon a leaf that symbolizes healing. Initially, I wanted to show a comparison between the public and the private sector. To do this I divided the form into two halves. Eventually, I decided that it didn’t help to support the larger narrative so I scrapped this idea, too. Finally, I landed on a more convincing form in [C], which was much both relatable and provided a visual cue about the topic. It occurred to me that a stethoscope was a useful representation that could directly convey my topic in one glance. The three-part composition of the stethoscope looked minimal and easy to read.
Breaking Down the Final Output
I refined the sketch on paper to develop my final output, sized to print on an A0 sheet (33.1" x 46.8").
The left side compares the states based on their public health components viz. Hospitals, Hospital Beds & Doctors. The Y-axis depicts the number of hospitals. I arranged the states according to their geographical zones on the
X-axis. The right side displays a legend to guide the reader. Each stethoscope represents a state. Starting from the bottom, the diaphragm displays the WHO-recommended quantity of patients to which a doctor should attend, i.e., 1,000 patients per doctor. The red ripples around it reflect the actual number of patients to which a doctor attends. The length of the tube describes the availability of hospitals per 100,000 population. The size of the binaural tubes in the top part represents the number of hospital beds per 1,000 population. The progressive lines that build-up the shape of the binaural tubes in the top part provides a guide to the reader for counting the number of hospital beds manually (1 line=0.1 bed per 1000 population).
I concluded the visualization by drawing the reader’s attention to elements to encourage them to form insights. One such insight is that doctors are seeing far more patients than is recommended. Another suggests that India is holding its own in terms of hospital bed capacity compared to most COVID-19-affected countries. And, finally, the identification of states that form 70% of India’s population and lie below the national value of hospital beds.
This was my first attempt at working with a dataset of such scale. I want to thank the data visualization community, especially giorgia lupi and Federica Fragapane, for their supportiveness and transparency about their process.
Thank you for reading. Your comments are welcome.
Sujay Kolgaonkar is a Master’s of Design student at the National Institute Of Design, Bangalore, India. He is learning Data Visualization as part of a course in Information Design.
