Reviewer has chosen to be Anonymous

Summary of paper in a few sentences:

The paper provides a detailed description of the Social Sentiment Indices powered by X-Scores (SSIX) project which aims at compiling a set of tools used to analyse and
understand social media attitudes for any given topic ranging from finance and health to education and security.
The infrastructure is defined as a Lamba architecture due to the need for both batch and real-time processing. The project relies upon existing scalable open source text mining tools in order to extract sentiment signal from cross-lingual social datasets (social media, public media outlets). Several business case studies are presented to emphasize on the importance and usefulness of such a tool.

Reasons to accept:

Using sentiment signals extracted from social media in a variety of domains is undeniably an important tool for both research and industry.
While there exist a wide range of free or proprietary tools for sentiment extraction from social datasets, I do acknowledge the need for a single open-source infrastructure that provides sentiment indicators able to serve any given topic in any given domain.

Reasons to reject:

Although the overall architecture is defined, I found key points that are not considered:
There is no mention about the dictionaries used in sentiment analysis, which are both very important and also very domain-dependent. Positive and negative words vary a lot across considered domains (financial terms are very different from words used in health topics, for example).
Another concern that is not addressed is how will the different data streams be classified into the specific domains they refer to? How will we know that a certain conversation on Twitter talks about finance, or health or education? From what I gather, this is part of data enrichment process in the ingestion tier. However, this is not a trivial task and further consideration should be given.
Also the authors mention that "it is important to acquire the ability to discern between irrelevant rumours and significant trending topics". I do agree with the sentence, but I don't see how this will be provided by the platform. There's a vast rumour detection literature to be considered here.

The paper is not self contained, there are quite a few terms introduced or mentioned which are not explained or even referred to throughout the paper. For example, there are mentions about "internal data format" which is never defined or referred to.

Related work on sentiment analysis is not covered in a sufficient way. Also, the novel contributions the project plans to bring as compared to existing approaches, should be stated more clear.

Other remarks:

SSIX platform behaviours: "behavior" is not the most appropriate term. I recommend "functionalities".
The resolution for the figures needs to be improved.

Nanopublication comments:

Further comments:

Reviewer has chosen to be Anonymous

Summary of paper in a few sentences:

The paper “SSIX Big Data Technologies and Methods for Leveraging Social Sentiment Data in Multiple Business Domains” presents a research project which aims to provide tools for sentiment analysis of social media and news data to European SMEs. The project’s main focus is finance applications, where indices based on social sentiment will assist investment decision making.

The paper is an extension of the publication “Social Sentiment Indices Powered by X-scores” [1].

The main extensions are:
- A description of the SSIX architecture (Section 3)
- Two usage scenarios (Sections 4.1 and 4.2)
- A description of the concept of sentiment analysis, and a description of four software libraries which provide sentiment analysis functionality for SSIX (Section 6.2)
- An extension of the planned business case study presented in [1] and two additional planned business case studies. (Section 7)

[1] Davis, Brian, et al. "Social sentiment indices powered by X-scores." ALLDATA 2016 (2016): 21.

Reasons to accept:

- Overall, the paper is well written and the different aspects and steps of the research project are described in sufficient detail.

- The proposed system fills a gap regarding the availability of easily accessible and customizable tools for sentiment analysis of large scale social media and news data.

- The presented figures are intuitive and contribute greatly to readability of the paper and to a better understanding of the proposed project.

- The business case studies and usage scenarios demonstrate that the system has the potential to meet the business needs of SMEs.

Reasons to reject:

- In my opinion, the delta to [1] is not large enough to warrant a new publication. Large parts of the paper are nearly identical to [1] (including the introduction, related work, the description of SSIX templates, big social and news data management as well as significant parts of other sections).

- This paper was submitted as a research paper. However, it describes a planned software system and there are no experiments and no evaluation of the system. (While the authors state that the four sentiment analysis libraries described in Section 6.2 “were evaluated in a qualitative nature on a Twitter data sample”, neither the sample nor the process of evaluation is described.)

- It is unclear which parts of this system, if any, are already implemented.

Nanopublication comments:

Further comments:

Clarification on “Suggested decision: Undecided”: The decision whether this manuscript should be accepted for publication should be based largely on the journal’s policy on
- how much accepted papers should differ from existing publications, and
- whether a description of a planned software system should be accepted as a research paper.

Clarification on “Technical Quality: Unable to judge”: No quantitative experiments or results are reported in the paper.

Below are some minor remarks on those sections of the paper which do not appear in [1]:
- Section 3 would benefit from a short introduction.
- The usage scenarios would benefit from additional references, e.g. studies where social media sentiment has been used for real-time polling.
- In Section 6.2, it should be clarified under which circumstances which sentiment analyzer should be used.
- Even though the business case studies are well presented, it would be interesting to know how the new set of tools provided by SSIX integrates with the existing infrastructure of the business partners.

Typos:
- Section 1: “The European research project Social Sentiment Indices powered by X-Scores
(SSIX), seeks to assist” -> remove comma
- Section 6: peopl’s -> people’s