Visual Data Story Protocol: Internal Communications from Domain

Expertise to Narrative Visualization Implementation

Apiwan Duangphummet

and Puripant Ruchikachorn

1,2 b

Institute of Public Policy and Development, Bangkok, Thailand

Department of Statistics, Chulalongkorn University, Bangkok, Thailand

Keywords: Data Storytelling, Visualization Design, Visualization Development.

Abstract: Data stories play an important role in effectively and intuitively communicating data insights as well as

enabling the audience to understand important social issues. Crafting a data story needs several sets of skills,

we propose a five-phase data story protocol in order to guide data story design and development, and promote

interdisciplinary team collaboration. This protocol was developed from our working team reflection on four

data story projects and researching the related work. We hope that this protocol could be one potential way

for non-journalism organizations to conduct data stories for their target audience.

1 INTRODUCTION

In recent years, data visualization has become

popular for its ability to transform data, information,

and knowledge into a form that relies on the human

visual system. Many news organizations especially

online journalists have been incorporating data

visualization into their narratives, often called ‘data

story’ (E. Segel and J. Heer, 2010).

Creating a data story is challenging and requires

interdisciplinary collaboration. Many visualization

models and frameworks have been proposed to

guide the design and development, mostly for

visualization systems. We therefore took lessons

learned from our experiences working on

communicating public issues through data, and a

careful analysis of related work on visualization

process and visualization collaboration in order to

propose a data story protocol consisting of five

major phases. While many frameworks for

visualization creation are developed by journalists

or information visualization specialists, we hope

that our work offers practical protocol created by

practitioners in other fields.

https://orcid.org/0000-0002-8080-9402

https://orcid.org/0000-0002-2721-6915

2 RELATED WORK

We conducted a literature review in the two main

topics: data visualization design and development

process, and data visualization design and

development collaboration.

2.1 Data Visualization Design and

Development Process

Stuart K. Card et al.

(

1999) suggest a four-stage

information visualization reference model: raw data,

data tables, visual abstractions and views. The first

two stages of the model are similar to data preparation

of our data story protocol. The last two stages of the

model are similar to visualization design and

visualization development of our protocol. The

difference is that our data story protocol requires

conceptualization for team alignment and realization

for insight discovery while Stuart K. Card et al.’s

model for information visualization development

does not require these phases.

Fry (2007) presents seven stages of data

visualization workflow: acquiring, parsing, filtering,

mining, representing, refining and interacting. More

than half of the processes have to do with getting data,

preparing data and making a visualization work

240

Duangphummet, A. and Ruchikachorn, P.

Visual Data Story Protocol: Internal Communications from Domain Expertise to Narrative Visualization Implementation.

DOI: 10.5220/0010327602400247

In Proceedings of the 16th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP 2021) - Volume 3: IVAPP, pages

240-247

ISBN: 978-989-758-488-6

properly. Our data story protocol emphasizes these

phases too in the data preparation, realization and

visualization design.

Munzner (2009) presents a nested model for

visualization design and validation with four tasks:

problem characterization in the vocabulary of the

problem domain, data and operation abstraction

design, visual encoding and interaction technique

design, and algorithm design to execute techniques

efficiently. This model has made a contribution to a

broad range of visualization papers, including design

studies, visualization technique study, frameworks,

and systems. Although the nested model guides the

design and validation of visualization systems and our

protocol guides the design and development of data

stories, both have similar core processes. The first

level to the last level of the nested model are similar

to conceptualization, data preparation, visualization

design and visualization development of our data

story protocol, respectively. Our protocol for data

story needs a realization phase to explore useful

insights while Munzner’s model for visualization

does not need this phase.

Design study approach is used for conducting

problem-driven visualization research in an

application area. Sedlmair et al. (2012) propose a

methodological framework for conducting design

studies consisting of the three following phases: a

precondition phase, a core phase and an analysis

phase. Both Sedlmair et al’s model and our protocol

are designed to be practical guidance on collaborating

with domain experts. The difference is that our data

story protocol focuses on the core phase of Sedlmer

et al.’s framework as we assume that learning

visualization literature, selecting promising

collaborations, and identifying collaborator roles in

the precondition phase are prepared and a data story

creation does not need the analysis phase to reflect

and write a design study paper.

McKenna et al. (2014) present the design activity

framework based on Munzner’s nested model. This

model consists of four tasks: understanding the

problem domain and target users, idea generation, idea

concretization into tangible prototypes and prototype

deployment. Walny et al. (2019) present five major

phases of their design projects: project

conceptualization, data characterization, visualization

design, visualization development, deployment and

use. Although McKenna et al.'s and Walny et al’s share

common phases with our data story protocol, both

works do not focus on data preparation.

Nina McCurdy et al. (2016) propose the four-

stage Action Design Research (ADR) methodology:

problem formulation, building, intervention and

evaluation, reflection and learning, and formalization

of learning. Both the ADR methodology and our data

story protocol are designed to tackle real-world

problems and share common stages. However, the

ADR methodology focuses on intervention and

learning for visual analytics systems research while

our data story protocol focuses on visualization

design and development processes for

communication.

In summary, our data story protocol partially

shares common phases with several data

visualization creation models and frameworks. The

difference lies in that many frameworks are

developed by journalists or information

visualization specialists, our work offers a protocol

created by practitioners in public policy in

collaboration with domain experts from different

fields. We hope this protocol can be a generalized

guideline for any practitioners who desire to

communicate important issues through data stories.

2.2 Data Visualization Design and

Development Collaboration

Visualization viewer background is not limited to the

sciences and engineering, but also other fields:

economics, business, and humanities, for example.

The knowledge and expertise for visualization design

and development are no longer restricted to computer

science. Visualization expertise requires several sets

of skills including human-centered design,

evaluation, cognition and perception (Kirby and

Meyer, 2013).

McCormick et al. (1987) present five types of

interdisciplinary team members: computational

scientists and engineers, visualization scientists and

engineers, system support personnel, artists and

cognitive scientists. Kirby and Meyer (2013) further

proposes an updated list of the roles: domain experts,

visualization experts, designer and human-computer-

interaction experts, cognitive and perceptual

psychologists, data analysis experts, database and data

management experts, and high-performance and high-

throughput computing experts. The first two roles are

the primary ones. Each of the other roles is either

assigned to a team member or assumed by a primary

member. Real-world visualization design projects

often consist of team members with diverse and

overlapping subsets of these skills. Our data story

protocol is closely aligned with the latter work in which

domain experts and visualization experts are involved.

3 DATA STORIES AND

PROTOCOL DEVELOPMENT

In this digital era, the broader accessibility of data

has dramatically increased the quantity of

Visual Data Story Protocol: Internal Communications from Domain Expertise to Narrative Visualization Implementation

241

information directed toward people including policy

makers and the general public. At the Institute of

Public Policy and Development (IPPD), a policy

laboratory and impact-oriented platform, we try to

utilize data story as one of effective ways to present

idea, intelligence, evidence, and public opinion in a

visual format for long-term sustainable

development.

Regarding overall storytelling format, we chose

the scroll-based and long-form of articles widely used

in journalism called “scrollytelling” (Doris Seyser

and Michael Zeiller, 2018) because this storytelling

format is designed to fit consumer behavior in the

digital world.

We have developed four data stories for IPPD

since 2019. As well as the literature review, the

reflections from our team members have guided our

data story protocol presented in Section 4.

3.1 Where is Thailand?:

Labor Productivity

The first data story (IPPD, 2020a) started from the

difficulty to gain insights from our visualization

shown in Figure 1. The visualization is a connected

scatterplot which can show the time series of two

variables at the same time (Steve Haroz, Robert

Kosara, Steven L. Franconeri, 2015). We gathered

several time series from various sources and hoped

a user can connect two previously separate variables

to comprehend the current state of Thailand—hence

a rhetorical question, Where is Thailand?—

compared to other countries and continent-level

aggregates. Despite a recent user study on connected

scatterplot (Steve Haroz, Robert Kosara, Steven L.

Franconeri, 2015), many early testers said that they

could not extract useful information from our

visualization.

We developed a data story to help users

understand and make sense of the same information

provided in the visualization. From our own

exploration, we found an interesting variable pair,

work hours per person per year and GDP per capita,

that could tell a story about labor productivity.

The data story starts with a controversial

question whether Thai people are hard working or

not. If the user scrolls down, the question fades out

as the visualization starts showing a line chart

(Figure 2) and a connected scatterplot of the related

global indexes to answer the question. As the user

scrolls through the presentation, the key

visualization maintains a consistent format and

changes only the content within the text box and

indexes.

What we have learned from developing our first

data story is that storytelling allows data visualization

to reveal analysis results compellingly and effectively

as we can see from increasing user questions and

opinions about Thailand’s labor productivity on

social media.

However, we spent four months on this data story

and wondered if the entire process of data story

design and development can take less time. Thus, we

started to reflect with the working team and

developed our first version of data story protocol

consisting of visualization tool development,

realization and presentation development, shown in

sketches in Figure 3.

Figure 1: A connected scatterplot of work hours per person

per year (horizontal axis) and GDP per capita (vertical

axis). Each circle represents a country with its flag or a

continent with its acronym. In this chart, Thai people

(bottom right) work long hours but do not generate much

GDP, compared to European counterparts (top left), for

example.

Figure 2: A data story about labor productivity. The key

visualization was based on the same information provided

in the visualization in Figure 1. The horizontal axis

represents time whereas the vertical axis represents annual

hours worked.

IVAPP 2021 - 12th International Conference on Information Visualization Theory and Applications

242

Figure 3: Our first version of data story protocol,

created from the working team reflection, has three phases:

visualization tool development for data analysis,

realization, and presentation development.

3.2 Where is Thailand?:

Plastic Management

The second data story (IPPD, 2020b) started from the

need to communicate an urgent issue about plastic

management to the public. It was not based on the

same visualization (Figure 1) as the first data story.

The data story on plastic management starts with

the prediction of waste in 2050 (Roland Geyer, Jenna

R. Jambeck and Kara Lavender Law, 2017). When

the user scrolls down, the prediction fades out and

another visualization starts showing with explanation.

Similar to the first data story, key visualizations stay

in the same formats, namely a line chart and a world

map (Figure 4).

Figure 4: Visualizations in the data story about plastic

management. The line chart on the left shows historical data

of cumulative plastic waste generation and disposal from

1950 to 2015, and projections of historical trends to 2050.

The world map on the right shows countries that have

introduced regulations on plastic bags and polystyrene

foam products.

Using our first version of data story protocol, the

entire process took only two months which is fifty

percent of time spent on the first data story. Other

observations for improvement are summarized as

follow:

1. The involvement of domain experts in the

working team improved the depth and variety of

content.

2. Not every data story required visualization tool

development for data analysis.

3. Realization phase took the longest time as we

changed the content direction multiple times. The

direction changes were mainly caused by the lack of

data.

We then researched related work and updated our

protocol to solve the mentioned issues. There were

five phases in our updated protocol as follows:

conceptualization, data preparation, realization,

visualization design and visualization development.

3.3 When Big Data Meets Small

Particles

Another interesting issue for data story development

was the exacerbating air pollution problem in the

greater Bangkok and the northern part of Thailand.

We used a mix of visualization techniques including

a stacked bar chart and various forms of geographical

maps.

Our third data story (IPPD, 2020c) starts with

important statistics about negative impacts of air

pollution in Thailand. As the user scrolls down, he or

she can interact with various interactive

visualizations as shown in Figure 5.

Figure 5: Visualizations in the data story about the air

pollution problem. First, the choropleth in the top left

compares Bangkok air quality index before and after

COVID-19 lockdown. Second, the choropleth in the middle

compares air quality index in the northern part

of Thailand before and after COVID-19 lockdown. Lastly,

the choropleth in the bottom right shows air quality in

Thailand throughout the reference year.

Using our current five-phase protocol, we spent

only two months creating the third data story. Another

key success factor was the involvement of domain

Visual Data Story Protocol: Internal Communications from Domain Expertise to Narrative Visualization Implementation

243

experts with strong analytical skills. In the making of

this data story, we collaborated with a data scientist

and a data analyst from a data analytics consulting

company. They primarily involved in the first three

phases: conceptualization, data preparation and

realization.

3.4 Reflections on Policy Options

for Road Safety

Thailand is one of the top five countries in the world

with the highest road traffic fatality rate (The World

Health Organization, 2018). So we picked this topic

for our fourth data story (IPPD, 2020d).

The data story starts with important statistics

about car accidents in Thailand. As the user scrolls

down, he or she can interact with different interactive

visualizations as shown in Figure 6. Similar to the

third data story, we used a mix of visualization

techniques including a bump chart and various forms

of bar charts.

We have learned that not every key visualization

needed to be interactive. We chose static format for

some key visualizations. Designing and developing

an interactive visualization usually takes more effort

than a static visualization, we should carefully

consider which key visualization is worth our team

effort.

Figure 6: Visualizations in the data story about policy

options for road safety. The bump chart on the left shows

the number of mortality rates by causes of death in 2018.

The bar charts on the right compare costs and benefits of

different policy options.

4 DATA STORY PROTOCOL

As data story design and development requires

several sets of literacy and skills, clarifying roles and

responsibilities (Table 1) of a cross-functional team

enables the teams to work efficiently and reduce the

unnecessary duplication of tasks.

Table 1: Involvement of each role in the protocol. Domain

experts are responsible for conceptualization, data

preparation and realization. Visualization experts oversee

visualization design. Visualization developers are in charge

of visualization development.

Phase in data story protocol

Role 1 2 3 4 5

Domain experts

✓ ✓ ✓

Visualization experts

✓ ✓

Visualization developers

✓

In practice, the working team could be divided

into three main positions: domain experts,

visualization experts and visualization developers.

Domain experts are typically researchers with subject

matter knowledge. Visualization experts are

specialists in effectively encoding data visually and

storytelling with data. Visualization developers are

software engineers with skills in graphic

representation creation.

The design objectives of our data story protocol

are to serve as a useful guideline for data story design

and development projects and to encourage

interdisciplinary collaboration between domain

experts. We illustrate the five phases with

documented artifacts from our last data story,

Reflections on Policy Options for Road Safety, as

follows. Please note that we do not describe data

analysis and visualization validation in details as

many previous works already covered the topics.

Table 2: A result from the conceptualization phase for

the data story ‘Reflections on Policy Options for Road

Safety’.

Target audience Channel Key messages

- General audience

who are interested in

or have knowledge

about wellbeing,

health and risk. An

interest in related

public policy would

be a plus.

- General audience

with curiosity to

learn from data and

understanding of

basic statistics

- IPPD

website

- IPPD

Facebook

page

- Road accidents are

the leading cause of

death in Thailand

comparing to other

causes

- The main risk

factors for road are

night-time driving

and motorcycle

driving

- Exploring what are

effective policy

options to reduce

road traffic fatalities

4.1 Conceptualization

As we begin working together as a team, it is

important that everyone on the team has a clear

understanding of project conceptualization. Domain

experts are responsible for identifying characteristics

of the target audience, distribution channel that we

IVAPP 2021 - 12th International Conference on Information Visualization Theory and Applications

244

expect the target audience to see our work and key

messages we would like to tell the audience.

(Table 2)

4.2 Data Preparation

After the scope of content has been defined, the

domain experts obtain and organize the data. This can

be either complicated (i.e., a dataset from an external

organization that requires a memorandum of

understanding) or very simple (i.e., readily machine-

readable dataset available in a public website).

A deliverable of this phase is the data that is relevant

to the use. A data dictionary is an optional

deliverable.

4.3 Realization

After having necessary data, the domain experts

discover useful insights and develop a storyline with

supervision from visualization experts. This step

involves basic statistics, data mining, and

storytelling. Oftentimes, defining key messages in the

first phase, data preparation in the second phase and

developing a storyline in the third phase are iterative.

For example, we set marine plastic leakage as one of

the key messages for the second data story ‘Where is

Thailand?: Plastic Management’ during the

conceptualization phase, but decided to cut off this

part and shifted focus to types of national policies on

plastic management during the realization phase.

A deliverable of this phase is the storyline that

includes detailed content and initial form of key

visualizations. An initial form of key visualization

does not need any detailed design. It could be

visualization from the original source or visualization

created by off-the-shelf software as we can see in

Figure 7.

4.4 Visualization Design

To quickly get feedback on the key visualizations,

visualization experts redesign the visualizations and

create visualization prototypes. An example is shown

in Figure 8. This step requires visualization expertise

to visually encode data effectively. Deliverables of

this phase are the final visualization prototypes.

In Figure 9, the change was made during the

visualization design phase. Visualization experts

suggested a line chart (Figure 8) representing benefits

and costs of every policy option scenario. However,

domain experts figured out data limitations and were

able to analyze only some scenarios. After a few

discussions, they came up with bar charts in Figure 9

to explicitly compare the benefit and cost of each

possible scenario.

Figure 7:Two key visualizations in an initial form. Treemap

on the top shows the number of road traffic deaths by type

of vehicle. Bar chart on the bottom shows the number of

road traffic deaths by time of day. These initial

visualizations were simply made by a data analyst and do

not need any design skills.

Figure 8: A visualization prototype of our last data story

‘Reflections on Policy Options for Road Safety’. In this

prototype, we tried to answer the question ‘which is the best

policy option on road safety?’ taking benefits and costs into

consideration.

4.5 Visualization Development

After finalizing key visualization prototypes,

visualization developers then define technical

requirements, develop and deploy the key

visualizations to support target devices. A deliverable

of this phase is the finished data story (see Figure 9)

that is ready for publishing.

Visual Data Story Protocol: Internal Communications from Domain Expertise to Narrative Visualization Implementation

245

Figure 9: One of final visualizations from our last data story

‘Reflections on Policy Options for Road Safety’. Starting

from the first prototype in Figure 8 and a few feedback

loops, we came up with an interactive visualization that a

user can compare policy options from five columns on the

left with benefits (blue bar charts) and costs (red bar charts)

on the right.

5 PROTOCOL USABILITY

TESTING

To evaluate the effectiveness, the efficiency and

satisfaction of the protocol, we used a modified

System Usability Scale (SUS) questionnaire

(J. Brooke, 1996). Widely used by many researchers,

SUS can quickly and easily collect a user's subjective

rating and can be used on small sample sizes

(A. Bangor, P. T. Kortum, and J. T. Miller, 2008). We

have also modified the SUS questionnaire to make it

suitable for the protocol by changing the word

‘system’ with ‘protocol’, and ‘functions’ with ‘parts’

as follow:

1. I think that I would like to use this protocol

frequently

2. I found the protocol unnecessarily complex

3. I thought the protocol was easy to use

4. I think that I would need the support of a

technical person to be able to use this

protocol

5. I found the various parts in this protocol were

well integrated

6. I thought there was too much inconsistency in

this protocol

7. I would imagine that most people would learn

to use this protocol very quickly

8. I found the protocol very cumbersome to use

9. I felt very confident using the protocol

10. I needed to learn a lot of things before I could

get going with this protocol

Figure 10: Box plot of SUS scores. The maximum score is

82.5, the average score is 60, the median score is 58.75 and

the minimum score is 47.5.

We surveyed six domain experts, visualization

experts, and visualization developers, who had

experience with our data story protocol. The average

SUS score was 60 (Figure 10). We compared SUS

scores of our protocol with nearly a thousand SUS

surveys for relative judgement. The protocol has

demonstrated an “okay” level of usability based on

the adjective scale presented by Bangor et al. (2009).

Table 3: Total weeks spent on data story creation and

the numbers of domain experts, visualization experts and

visualization developers involved in each data story.

Data

story

Weeks Domain

experts

Visualization

experts

Visualization

developers

116 0 1 1

28 5 1 1

36 7 1 1

46 8 1 1

6 DISCUSSION AND

CONCLUSIONS

We have presented a five-phase data story protocol

formed by reflecting on our experiences as members

of the working team for four data stories and the

literature review.

Regarding a usability testing, the protocol

usability is in the marginally acceptable area. That

means we need to improve the protocol in order to

increase the effectiveness, the efficiency and

satisfaction of the protocol. In terms of effectiveness,

we also compare total weeks spent and the number of

each role participated in each data story as shown in

Table 3. Using our protocol and involvement of

domain experts in the fourth data story can reduce

62.5 percent of total weeks spent, compared to total

weeks spent on the first data story.

Interesting areas for improvement came from the

user feedback including brainstorming and early

involvement of an approver. Some domain experts

said that it was difficult to come up with ideas on key

visualizations. Thus, we suggested adding a

brainstorming session between domain experts and

visualization experts as a part of the realization phase.

Some visualization experts also mentioned that they

had to revise the entire data story that had been

IVAPP 2021 - 12th International Conference on Information Visualization Theory and Applications

246

developed due to late feedback. An early approval

from the decision maker of a data story project should

be added, at least for the deliverables of the

conceptualization, realization and visualization

development.

We generalize data story design and development

processes with the hope that this protocol could be an

alternative approach for the practitioners in the non-

journalism industry to effectively communicate vital

issues for improvement through data stories.

ACKNOWLEDGEMENTS

We would like to record our gratitude to the protocol

users for their feedback and dedication in data story

creation, and visualization experts who have been a

part of this research for their insightful guidance and

support.

REFERENCES

Bangor, A., Kortum, P. T., & Miller, J. T. (2008). An

Empirical Evaluation of the System Usability Scale.

International Journal of Human-Computer Interaction,

24(6), PP.574-594.

Bangor, A., Kortum, P. T., & Miller, J. T. (2009).

Determining What Individual SUS Scores Mean:

Adding an Adjective Rating Scale. Journal of Usability

Studies, 4, PP.114-123.

Brooke, J. (1996). SUS: A “Quick and Dirty” Usability

Scale. Usability Evaluation in Industry, PP.189-194.

Taylor & Francis, London.

Card, S., Mackinlay, J., Shneiderman, B. (1999). Readings

in Information Visualization: Using Vision to Think.

Morgan Kaufmann Publishers, San Francisco.

Fry, B. (2007). Visualizing Data. O'Reilly Media, Inc.

Geyer, R., Jambeck, J.R. and Lavender Law, K. (2017).

Production, Use, and Fate of All Plastics Ever Made.

Science Advances, Vol.3, No.7, e1700782.

DOI: 10.1126/sciadv.1700782.

Haroz, S., Kosara, R., Franconeri, S.L. (2015). The

Connected Scatterplot for Presenting Paired Time

Series. IEEE Transactions on Visualization and

Computer Graphics. 22. PP.1-1.

DOI: 10.1109/TVCG.2015.2502587.

Institute of Public Policy and Development (IPPD).

(2020a). Where is Thailand?: Labor Productivity.

Retrieved from https://ippd.or.th/thailand-productivity-

story.

Institute of Public Policy and Development (IPPD).

(2020b). Where is Thailand?: Plastic Management.

Retrieved from http://www.data.ippd.or.th/plastic-

story.

Institute of Public Policy and Development (IPPD).

(2020c). When Big Data Meets Small Particles.

Retrieved from http://www.data.ippd.or.th/plastic-

story.

Institute of Public Policy and Development (IPPD).

(2020d). Reflections on Policy Options for Road

Safety. Retrieved from

https://data.ippd.or.th/roadsafety.

Kirby, R.M. and Meyer, M. (2013). Visualization

Collaborations: What Works and Why. IEEE Computer

Graphics and Applications, 33(6),

PP.82–88. DOI:10.1109/MCG.2013.101.

M. Sedlmair, M. D. Meyer, and T. Munzner. (2012). Design

Study Methodology: Reflections from the Trenches and

the Stacks. IEEE Transactions on Visualization and

Computer Graphics, 18(12), PP.2431–2440.

McCurdy, N., Dykes, J. and Meyer, M. (2016). Action

Design Research and Visualization Design.

In Proceedings of the Sixth Workshop on Beyond Time

and Errors on Novel Evaluation Methods for

Visualization (BELIV '16). Association for Computing

Machinery, New York, USA, 10–18.

DOI: https://doi.org/10.1145/2993901.2993916.

McCormick, B.H., DeFanti, T.A. and Brown, M.D. (1987).

Visualization in Scientific Computing. Computer

Graphics, Vol. 21, No. 6.

McKenna, S., Mazur, D., Agutter, J. and Meyer, M. (2014).

Design Activity Framework for Visualization Design.

IEEE Transactions on Visualization and Computer

Graphics, 20(12), PP.2191–2200.

Munzner, T. (2009). A Nested Model for Visualization

Design and Validation. IEEE Transactions on

Visualization and Computer Graphics, Vol. 15, No. 6,

PP.921-928, DOI:10.1109/TVCG.2009.111.

Segel, E. and Heer, J. (2010). Narrative Visualization:

Telling Stories with Data. IEEE Transactions on

Visualization and Computer Graphics, Vol. 16, No. 6,

PP.1139-1148, DOI:10.1109/TVCG.2010.179.

Seyser, D. and Zeiller, M. (2018). Scrollytelling – An

Analysis of Visual Storytelling in Online Journalism.

International Conference on Information Visualisation

(IV). 401-406. 10. DOI: 1109/iV.2018.00075.

The World Health Organization. (2018). Global Status

Report on Road Safety 2018. Retrieved from

https://apps.who.int/iris/bitstream/handle/10665/27646

2/9789241565684-eng.pdf.

Walny, J., Frisson, C., West, M., Kosminsky, D.,Knudsen,

S., Carpendale, S., Willett, W. (2019). Data Changes

Everything: Challenges and Opportunities in Data

Visualization Design Handoff. IEEE Transactions

on Visualization and Computer Graphics. 1-1.

DOI: 10.1109/TVCG. 2019.2934538.

Visual Data Story Protocol: Internal Communications from Domain Expertise to Narrative Visualization Implementation

247