By Rafael Martins https://orcid.org/0000-0002-2901-935X

Review Details

---

Reviewer has chosen not to be Anonymous

Summary of paper in a few sentences:

The paper defends the position that visualization techniques should be valued as important scientific artifacts in the field of data science. It does that by exploring, in general terms, what is a visualization, how visualizations are designed, what is their importance, and how they can help with the exploration of large amounts of complex scientific data. The author uses omics data visualization as an example of knowledge domain where visualization techniques can be useful tools for scientists.

Reasons to accept:

The goal of the paper is to defend a position, as described above (that visualizations are important scientific artifacts) and it does that well. It is well supported by good references, it explores the subject in a broad yet objective manner, and succeeds to communicate its point. It is very well written and easy to read and understand. Finally, it fits well in its role as editorial / position paper, introducing a concept without going too deep in specific problems and challenges. For this reason, I believe that the balance is in favor of the paper's advantages, so I recommend the acceptance of the paper.

Reasons to reject:

I believe there are no strong reasons to reject the paper. As a disadvantage, I'd mention that I believe that some parts could have been better explored. The paper mentions that the visualization of omics data is at its core, but very few concrete examples were described, and even those in very few details. I feel that, if possible, the author should have described one or two concrete cases of omics visualization in more details, in order to make the paper's position stronger. Most of the text is developed with general visualization in mind.

Nanopublication comments:

Further comments:

Reasonable background:
For being a position paper/editorial, I understand that the paper is not supposed to cite the absolute state-of-the-art works on the area, but instead it should use a good set of examples. In that case the paper succeeds. Still, I was left with an overall feeling that the author is not deeply acquainted with the publication venues for visualization, specially when she mentions that "...visualization research can be overlooked and not interpreted as a valuable publishable scientific effort." Maybe she means that, for this specific venue (Data Science), that is the case; if that is so, I can understand the statement and agree with it.

Limited novelty:
It is a position paper/editorial, so I expect that it is not supposed to be novel.

By Alexander Lex https://orcid.org/0000-0001-6930-5468

Review Details

---

Reviewer has chosen not to be Anonymous

Summary of paper in a few sentences:

This paper argues the need for visualization when working with omics data.

Reasons to accept:

Overall, I think it's important to publish a paper like this in Data Science.

Reasons to reject:

I think this paper could be much improved by teaming up with a biology visualization expert and bringing the recommendations more in line with the current state of the art in visualization.

Nanopublication comments:

Further comments:

While I applaud the general direction, I find the paper lacking in some regards. Some of the arguments seem ad-hoc. For example, the "Dangers when creating visualizations" section discusses one or two challenges one needs to overcome, but clearly isn't comprehensive. I'm unsure why the problem of biases is specific to visualization; rather I believe it to apply to all data analysis methods. Other challenges aren't mentioned.

The author mentions (and implicitly recommends) visualization techniques that, while they are published and used, don't align with commonly agreed-upon best practices in visualization. Examples are 3D node-link visualizations, rose diagrams, star and glyph plots, map and landscape views. These might have been mentioned in outdated literature, such as the referenced "Readings in Information Visualization", which is an excellent book, but has to be taken with a grain of salt nowadays, as the field has matured significantly in the last 18 years.

For the process of building visualizations targeting domain users's needs, I'd recommend to include a reference to and insights from the design study methodology paper [1].

[1] Sedlmair, Michael, Miriah Meyer, and Tamara Munzner. "Design study methodology: Reflections from the trenches and the stacks." IEEE transactions on visualization and computer graphics 18.12 (2012): 2431-2440.

Reviewer has chosen to be Anonymous

Summary of paper in a few sentences:

This paper positions the use of visualizations within the field of omics technologies. It discusses the role visualizations can have in clarifying data within the biomedics and the dangers it can lead to when interpreting data. It posits a more specific role for visualization researchers in relation to complex data sets.

Reasons to accept:

no reason to accept

Reasons to reject:

This paper does not address a novel problem within the visualizations research. It posits that data in the omic technologies is quite complex to be able to visualize in correct manner. This is not a novel issue or specific type of data. Also, the author does not explicitly clarify the novelty of the issue addressed.
This position the paper takes is not quite clear, and the paper is quite repetitive, repeating the use and function of visualizations within biomedics.
The dangers of visualizations that the author points out are not well argued, while the introduction of the paper makes clear that it wants to point out the deficits of visualizations within the omics technologies.
The dangers described
The arguments on the role of visualizations researchers are not well substantiated.
The summary of the paper points out developments that have not been discussed in the paper

The structure of the paper is not clear and there are quite some language mistakes.
While the length is far beyond the maxim number of words, the short length might be the reason the paper misses substantiated arguments and clarifications.

Nanopublication comments:

Further comments: