Adding Vaccination Data | Deloitte Analytics and Deloitte Digital

Wireframes, Design and Data Visualization

🏆 Winner of 2022 Ministry of Justice’s Governmental Transparency Award 🏆

In December 2020, MOH decided to update and redesign the dashboard to track the effectiveness of the COVID-19 vaccination on Israeli morbidity over time. MOH’s approach was that as they’ll give access to more data at higher resolutions, and people will be able to ‘play’ themselves with data and make their cuts and tests - the more trust they will create.

Throughout the time, additional types of users were considered, as it became apparent that they were also viewing dashboard data: academic and public health researchers, local authorities, and journalists (more information could be found here).

The more information we reflect, the more trust we will create.
— Rona Kaiser, CIO at Ministry of Health Israel

The Problems

  1. The general public doesn’t know how to get relevant and accurate information about the vaccinations

  2. The stakeholders don’t have access to visual and digestible vaccination information

The Goals

  1. Draw attention to the vaccination project’s progress

  2. Establish trust and transparency between MOH and the public by sharing reliable information

  3. Provide the public with information that will assist in minimized risk-taking and informed decision-making

  4. Motivate the public’s willingness to get vaccinated, and stay safe 

  5. Support policy-makers and other stakeholders by continuously providing up-to-date information and progression regarding the virus's spread

The Process

The Current Dashboard (Before the Changes)

First Version - Adding Three Vaccination Metrics + KPI

A decision was made: three simple tiles are created, and each will reflect vaccination metrics.
The vaccination metrics tiles will replace the virus-spread progression tiles.

We aimed to reflect the following metrics:

  • The daily number of vaccinated individuals

  • The cumulative number of vaccinated individuals

  • The total percentage of vaccinated individuals

We deliberated what is the appropriate granularity level of data for all dashboard stakeholders:

  1. Health maintenance organizations - we decided against displaying this distribution, considering possible variations in vaccination progression between HMOs (what might evoke a negative sentiment towards HMOs who lag).

  2. Prioritized populations - e.g., at-risk groups or medical staff.
    First, it was considered to display the cumulative percentage of vaccinated individuals from these groups since the graph trendline is increasing fast. But since data segmentation between the general public and these groups was unclear, it was decided to leave it (and instead stick with the more valuable general-population overview). Once data segmentation is precise, it will be displayed at a lesser dominant part of the screen.

Changes and Tests done:

At the beginning of the vaccination project, showing an overview of the cumulative number of vaccinated individuals also posed a challenge. At the start, as the vaccination volume was starting to pile up, graphs still looked deflated. The solution was to wait a minimum of seven days before displaying this data.
The logic behind choosing the type of charts was - for a daily breakdown, a column chart, and an area chart or line chart for a cumulative view.

At the top part of the dashboard, under the KPI tiles, the test to display the vaccination metrics tiles instead of the spread metrics tiles can be viewed (among them, the tile showing the cumulative number of vaccinated medical staffs). However, it was decided against showing this tile, and we rolled back to the previous graph of the general population. 

At the left bottom of the dashboard, a vacancy was filled with critical and severe patients’ data (derived by the spread metric we’ve decided not to show anymore). The feedback received by MOH’s managers was that this data is repetitive, as a similar metric of severe-conditioned and intubated patients already exists.

Here we re-added the total percentage of vaccinated individuals instead of the percentage of vaccinated medical staffs. An additional update was to add a metric of total vaccinated individuals as a KPI tile. The idea behind it was to display an absolute number that is easily noticeable. After this, we were short of space at the KPI part. Therefore it was decided to keep the cumulative deceased KPI tile (and remove the intubated-patients KPI tile) since a deceased daily breakdown tile already existed below, and we wanted to create a connection between these parts.

Second Version - Replacing the Navigation and Making Small Improvements
In this version, I added the new navigation (more information could be found here) and recolored the Daily Vaccines Administered tile since we wanted to enrich the color palette. This main dark green color will come across at the Inpatient Beds Occupied Breakdown tile at the bottom of the dashboard, to the right.
The governmental plan was to start vaccinating the second dose on January 10th, so we also received a request to add the second dose of vaccination to its KPI tile.
This version also includes the content areas and the Traffic Light Model tiles.

Third Version - Adding the second Dose of Vaccination Data

In this third version, we’ve added the second dosage to each tile. Since a dark green color was added by us, it was possible to use dark green color to represent the first dosage, and the second booster shot was represented by the light green color. We transformed the current charts into stacked charts representing multiple data groups.

The range of the graphs was dynamically changed to fit the scale of vaccination numbers as well (i.e. it changes automatically, according to the data inserted).

Additionally, few visual improvements were made in favor of accuracy and technical reasons. For accuracy and distinction, We’ve exchanged rounded columns with sharp-edged ones. Also, in the vaccinated KPI tile, a thin line between the first and second time vaccinated was added. Technicalities drove us to remove the stripes and to leave only grayed information, using two shades of gray for different data points.

Demographics: Tiles of most-vaccinated municipalities and vaccinations by age-groups

At the beginning of January 2021, MOH was flooded with requests for further information. It became apparent that a different breakdown of data would better serve the public and the stakeholders in favor of transparency. We started working on a tile representing the leading municipalities regarding vaccination percentage and a second tile of vaccinations by age-groups.

  • The age-group tile (on the left) was designed several times until it reached its final design:

  1. In the beginning, the tiles were designed so that each vaccination dosage would have its designated tile, displaying the percentage of vaccination by age-group.

  2. Later on, we unified both tiles (1+2 dosages) to one tile that shows both dosages - in which a user can filter data by the relevant metric.

  3. Afterward, the design was changed to a stacked bar chart, in which the percentage of vaccinations in both dosages could be seen.

  4. In the final version, the order of charts was switched to display the second dosage on top of the first dosage. This created an issue since further dosage-specific information would only be displayed by a tooltip if users hover over the relevant bar. In addition, it was unclear where the count of the bottom bars starts (beginning of the y-axis or starting from the end of the top bars). It was challenging to understand the percentages of vaccinations for both dosages since they’re not displayed one next to the other. This is an example of a compromise driven by limited screen space and choosing a compact design for our charts.

  • The most-vaccinated municipalities tile (on the right) was designed similarly to the Traffic Light Model charts (including displaying some of their metrics and the daily score). This was done this way since we still didn’t have important information to display for municipalities and vaccination percentages. When a space opened up on the screen, this tile was expanded vertically to fit two columns. Therefore we’ve decided to add the vaccination percentages’ progress bar for both dosages besides the numbers-based display. Thus it was visually clearer to track the progress on each metric.

Age-related data dilemmas

  • Later on, we deliberated on how to display age-related data. We wanted to show data for 10-19 year-old vaccinations, while data was available only for 16-19 year-old. Inquiries received by MOH’s spokesman said it’s misleading to display data labeled “10-19” if younger children couldn’t get vaccination yet. After further discussions, the age-group remained 10-19, in accordance with official publications of the state of Israel.

    In 2021’s summer, vaccinations for the age-groups of 12-16 started, ranges were changed to 12-15 and 16-19, and data for the third dosage was added.

  • Since at this stage, there were no vaccinations for babies (ages 0 onwards), this part (“0-9”) was eventually taken out (numbers shown here are for visualization purposes only and don’t reflect actual data).

Mobile version - most-vaccinated municipalities tile

Adaptation of the table for mobile by creating alternating columns: I designed two states of the table. After pressing the “active patients/other metrics” button (respectively), the screen is switched to display an additional column on the left side of the table. This was designed this way to prevent the users from horizontal scrolling.
Here I applied the same logic as in the mobile design of the Traffic Light Model tile.

The Final Dashboard

More Case Studies

01

Adding the Traffic Light Model Charts

02

Redesigning Dashboard 2.0: Creating Content Areas