Public Deliverables

WP#Deliverable AbstractDue Date
WP11.9Final ReportPlaceholder deliverable, no document will be available. As agreed with the Project Officer, this deliverable is a placeholder (as a document must be uploaded to the portal); to avoid duplication of effort for both the partners and the reviewers all content intended for this periodic report will be included in D1.9 - Final Report and the Year 3 - Technical Report Part B .28/02/2018
WP22.1Business Requirements and Business Cases' DefinitionsThis deliverable defines the business requirements industry partners have that will lead to the detailed construction of business case studies. A full usage description of how SSIX indexes will be implemented and used within a real world scenario will be provided in order to guide the development in WP4 and deployment and validation in WP5.31/08/2015
WP22.2Business Methodology DefinitionThis Deliverable describes the business methodology and business steps required by each partner for implementing and adopting the software developed by the SSIX consortium. While the software developed is intended to be as generic as possible, each industrial partner has very specific and unique commercial goals to pursue while using social sentiment analysis. Therefore, each partner has identified elements of their business process that are specific to their business needs which requires personalisation of SSIX platform.29/02/2016
WP33.1Data Requirement Analysis and Data Management Plan V1This document aims to provide a detailed overview of the platforms and techniques that can be used as data sources for the entire SSIX platform.
The document clearly lists all the public data available that can be retrieved and processed by the SSIX platform, along with the detailed results of the assessments performed on the identified data sources. This document will help to highlight important structural aspects of the platform and to identify all the criticalities that have to be taken into consideration when dealing with certain data collection techniques.
WP33.2Data Management Plan V2This deliverable contains the second version of the Data Management Plan, that is the document describing the data management lifecycle for all the data sets that will be collected, processed or generated by the SSIX project.31/08/2016
WP33.3Data Management Plan V3An updated version of the Data Management Plan (DMP), describing all the types of data that will be collected and stored during the project, how this should be recorded, stored and protected, in line with the ethical procedures and laws on data protection and what can be made publicly available for reuse by other parties through the Pilot on Open Research Data in Horizon 2020.28/02/2017
WP33.7Data Streams documentationThis deliverable contains the technical documentation of the Streaming APIs defined and implemented in WP3 - Data Management; needed to send data to the provided clients and retrieve data from the data ingestion infrastructure.
As the project is still in a development phase, the current target of this document are the technical partners of the consortium, which are connecting to the endpoint to source data for the NLP Pipeline in WP4 - and dashboard in WP5. In the future, once software releases of the SSIX platform are initiated, other actors may need to access the data APIs; hence the need to have an official technical documentation that can be consulted by future users of the whole infrastructure.
WP44.1NLP Service and Analysis Architecture (Initial Version)This document is an annex to the actual software deliverable. It describes the development of an initial architecture for the NLP Service and Analysis Pipeline for the SSIX Platform. This first iteration is based on existing and state-of-the-art technology, whereas the second iteration will take into account evaluation results and feedback from business cases to improve the stack.
The document is structured as follows: Section One introduces the scope and goals of the problem which the outcome of the D4.1 is going to solve in the SSIX project. In sections  2 & 3 processing paradigms and technologies are appraised as candidates to fulfil the target are presented and evaluated. Section Four describes the state-of-the-art technologies considered in this initial version to provide a solution on the sentiment analysis problem; which implementation details can be found in Section Five. Finally, Section Six closes the document with some conclusions and future lines to work for next milestones.
WP44.2NLP Service and Analysis Architecture (Revised Version)This document provides a revised version of the NLP Service and Analysis Pipeline architecture in the SSIX Platform. This second iteration supersedes the information of both D4.1 - NLP Service and Analysis Architecture (Initial Version) [Fernandez2015] and D4.5 - NLP Service and Analysis Pipeline (Proof-of- Concept) [Fernandez2016], gathering the current language resources [Capelletti2017a] and taking into account evaluation results and feedback from business cases to improve the stack.
The document is structured as follows: Section 1 introduces the deliverable and scopes the document. Then, Section 2 describes the architecture in details, including static and dynamic views, and its deployment. Section 3 presents the NLP features and components supported in this version. The internal API is described in Section 4. Section 5 describes some key infrastructures the architecture is relying on. Section 6 offers conclusions and presents future work.
WP44.31st Catalogue of SSIX Language ResourcesThis document presents a survey of the available language resources for Sentiment Analysis that are relevant to the SSIX project. It includes several kinds of resources (Corpora, Lexica, POS Taggers, Classifiers, etc.) for the main European languages (English, Italian, German, Spanish, French, and Portuguese)1. In general, the resources listed are intended to provide the main constituents of a Multilingual Sentiment Analysis platform.29/02/2016
WP44.42nd Catalogue of SSIX Language ResourcesThis catalogue includes the language resources that have been developed within the SSIX project.
These include annotated datasets, lexica, annotation tools, support for translation. Some l resources were under development during the drafting of this document. We report however those that we find relevant and for which a concrete plan of implementation is available.
The document is structured so that in Chapter 2 we discuss the various Gold Standard corpora that have been developed and the methods to develop Gold Standards for other languages via machine translation and text normalization. In Chapter 3 we provide an overview of the tools that we developed for data annotation and validation. Chapter 4 is about another language component which is essential to the SSIX project, named entity recognition. The sentiment classifier implemented for finance is described in Chapter 6.
WP44.6NLP Service and Analysis Pipeline (Prototype)Prototype implementation of the NLP Service and Analysis Pipeline, running on a scalable architecture and usable for evaluation in real- world scenarios.
After D4.5, this deliverable represents the second iteration over the SSIX NLP Pipeline, described in detail in D4.2.
WP55.1SSIX Process SpecificationThis document is addresses the main SSIX platform process flows and related business processes (in line with the business case studies) that are interlinked with SSIX software architecture. For this purpose the document presents the main elements to be taken into consideration when analysing the process flows, and then define the process flows to be expected in SSIX platform. In particular within the SSIX process flow there are identified where SSIX variables defined in SSIX Template can be updated in order to change SSIX platform behaviour as needed. Finally, conclusions are drawn for the benefit of the software development process.29/02/2016
WP55.2SSIX Architecture SpecificationThis document describes the SSIX system architecture. It consists of an introduction, a review of the business and software requirements, an introduction to big data and its architectures in light of literature, a high-level overview of the proposed system architecture and its adherence to big data architectures, a discussion on core tiers and their components, a section on development tools and techniques, a discussion of potential developments and a conclusion.29/02/2016
WP55.3SSIX Technical Validation PlanThis document describes a plan for evaluating the results from WorkPackage 4 - NLP Services and Analysis Pipeline (WP4) and WorkPackage 3 - Data Management (WP3) as integrated in the platform built within WorkPackage 5 - SSIX Platform Deployment, Validation and Evaluation (WP5) against the requirements tracked in WorkPackage 2 - Business Requirements, Use Cases and Business Methodology (WP2).
The document focuses on developing test cases and baselines to validate the implemented platform against the requirements from WP2 and measure the performance in the individual business use case studied in WP2. This is an open document that will be integrated throughout the project along with the formal definition of the SSIX platform resulting from WP3 and WP4 and WP5.
An Annex has been produced which contains confidential technical discussions and data shared between Consortium members, the purpose of which is to aid the development of SSIX technologies.
WP55.41st version SSIX Release PlatformSSIX aims to produce sentiment metrics from publicly available multilingual data sources to aid in understanding of aggregated or entity level sentiment and decision making. This document accompanies the first release of SSIX Platform and is meant to review its architecture and components and to serve as a guide for deploying SSIX at user’s premises or on Cloud. This document also discusses open issues and the roadmap of SSIX deployment up to its final release.28/02/2017
WP55.5SSIX Final Release PlatformThis document accompanies the final release of the SSIX platform. It updates existing documentation from the first release of the SSIX platform and describes the components, which have been developed further. It is technical documentation which serves as a reference for installation and deployment of the SSIX platform. This requires access to resources such as the source code repository and the SSIX project website, links to these are provided when needed.
It is divided into six sections and is focused on the deployment and functioning of the platform. The first section gives a brief overview of the SSIX project followed by a synopsis of the architecture for the development platform. The first section also contains an overview of the updates carried out during the final year of the project.
The second section begins by introducing the overall SSIX architecture, highlighting different layers and explains the operation of the different work packages within the SSIX platform. It then discusses the most appropriate deployment platform choices such as Docker Swarm1, Kubernetes2 and Datacenter Operating System3 (DC/OS). The presented information is updated to reflect the final status of the platform components and services.
Section 3 then lists all components present within the platform, along with necessary deployment information such as the location of important and relevant files within the Bitbucket code repository4, prerequisites and deployment parameters.
The fourth section explains the current environment of the platform deployment with available hardware resources and cluster information. It also describes deployment procedures and provides detailed information on the deployment. These can later be adopted for explaining deployment on other platforms such as DC/OS on Amazon Web Services (AWS)5.
Section five covers the way to add further X-Scores in the system. It includes source code snippets to let the developers know of appropriate packages and classes to use for adding custom X-Scores.
Finally, the sixth section gives a conclusion and discusses possible future directions for the platform.
WP55.6SSIX API DefinitionsThis deliverable contains the technical documentation of the REST API (Application Programming Interface) defined and implemented in WP5 - SSIX Platform; needed to send sentiment data to the clients. It will retrieve data from the Elasticsearch1 and the WP5 Engine. Any authorised user can use this API to get sentiment data.28/02/2017
WP66.1SSIX Technology and Skills Transfer ReportThis report is crucial for identifying the risks related to social media analysis in general and especially for SSIX from a stakeholders’ perspective within the context of technology and skills transfer across the consortium. Further, this report builds the basis for potential solution and mitigation strategies. Therefore, we surveyed the requirements and risks perceived by the market (i.e. potential users/ customers, management boards, other stakeholders in financial markets analysis, the SSIX use case partners and their customers). Additionally to this survey, we conducted several informal face-to-face interviews, for example, financial industry conferences (e.g. 'Battle of the Quants', 'Banken im Umbruch', 'Munich Finance Forum') with interested stakeholders in sentiment and financial markets analysis from all over the world. In addition to risks for the application of social sentiment analysis in finance, technological, ethical, and legal risks were also discussed.
Overall, no special risk was identified as the main concern for potential users of social sentiment analysis. Some risks that we had previously presumed as very important to the stakeholders are ultimately not that relevant to them as our survey and the conducted interviews reveal. Notably, and somewhat unexpected is that ethical risks do not play any considerable role for potential users. Technological risks play some role in the perception of the project's stakeholders but only to a minor extent.
The most important concerns expressed by our respondents are about the results of the SSIX metrics and the underlying datasets used in producing the results.
The most common questions include:
● Are there any biases or distortions?
● Does the indicator have a good predictability property?
● Are the results valid?
● Is the market coverage of the underlying data sufficient to draw results from it?
● Is the analysis or the underlying dataset influenced by some kind of self-fulfilling prophecy?
● Can the results be misinterpreted?
● How big is the volatility of the SSIX metrics?
We had been aware of most of these potential risks before and have included mitigation strategies in the use cases. Still, having the clear view from the potential end-users is of considerable benefit in making amendments and shifting focus where necessary when entering the market.
WP66.2Project Web site, Wiki, LinkedIn and Training Materials V1.0The purpose of this deliverable is to introduce and describe the website of the Project as one of the main instruments for dissemination. It describes its general outline and structure, the rationale and the type of information contents that can be found on the site with screen snapshots of some pages. This deliverable also outlines the various social media accounts the project will use to assist with dissemination. Finally this deliverable will layout the various planned training materials that are relevant to the project.30/04/2015
WP66.3Project Web site, Wiki, LinkedIn and Training Materials V2.0The purpose of this deliverable is to update and outline the current project website, social media presence and training materials for the SSIX project. The official deliverable due for submission is classed under (Websites, patents and Filling) and not a Report. Hence, this document acts as an addendum to the physical deliverable i.e. the project website, social media presence and training materials. This deliverable outlines the updates to the general shape and structure of the website, the rationale and the type of information content that can be found on the site with screenshots of some pages to help illustrate the changes. This deliverable also gives an update on the various social media accounts the project is using to assist with dissemination. Finally, this deliverable will give an update on the various training materials that are relevant to the project.29/02/2016
WP66.4Project Web site, Wiki, LinkedIn and Training Materials V3.0The purpose of this deliverable is to update and outline the current activities associated with the project website, social media presence and training materials for the SSIX project. The official deliverable due for submission is classed under (Websites, patents and Filling) and not a Report. Hence, this document acts as an addendum to the actual deliverable (D6.4 - Project Web site, Wiki, LinkedIn and Training Materials V3.0) i.e. the project website, social media presence and training materials produced during the project Workplan.
This deliverable outlines the updates to the general shape and structure of the website, the rationale and the type of information content that can be found on the site with screenshots of some pages to help illustrate the changes. It also gives an update on the various social media accounts the project is using to assist with dissemination.
Finally, this deliverable will give an update on the various training materials and workshops that are relevant to the project.
WP66.5Final Version of Project Web site, Wiki, LinkedIn and Training MaterialsDeliverable 6.5 - Final Version of Project Web site, Wiki, LinkedIn and Training Materials is a short document describing the activities concerned with the project website, social media presence and training activities for the SSIX project. The deliverable is classed under ‘Websites, patents and Filling’, hence, this document acts as a complement to the actual work carried out, i.e. the project website, social media activity and training activities delivered during the project.
This deliverable outlines the updates to the structure of the website, as well as a review of all the sections detailing the content that can be found with screenshots. It also gives an update on the various social media accounts the project is using to assist with dissemination. Finally, this deliverable will give an update on the multiple training materials/resources and training activities which the project has provided via the project Knowledge Base and training sessions.
WP66.6SSIX Professional skills and technologies training coursesThe purpose of this document is to provide an overview of training actions and resources related to the SSIX project. A number of resources have been developed. They fall in the domains of professional skills and technologies. The resources are planned to be used as training material for upcoming training sessions. All resources and training sessions will cater for two types of audiences, i.e., end users and technical experts in NLP and Machine Learning technologies. All these resources, as well as scientific publications, are available publically, and are located in different parts of the SSIX website.31/08/2017
WP66.7Technology Transfer and Dissemination Plan 1st VersionUpdated annually, the Technology Transfer and Dissemination strategy document contains a definition of the target groups, the activities, planned dissemination and training events that each partner is committed to carry out, the means to be adopted and the expected results of the dissemination. It will be drafted by the WP leader in the initial stage and later updated to provide Business Plan details. Planned dissemination activities will include:
1) Publications and Presentations: Frequent submission of articles retaken by the most important international journals and conferences in the field, as well as the participation to major international and European conferences and events.
2) Outreach: Practical sharing of project knowledge through sites (e.g. LexisNexis, PR Newswire, Global Risk Regulator, Online Traders Forum, CEB Financial Services Technology Summit, Big Data and Analytics for Financial Services, SemTechBiz Conference and Sentiment Analysis Symposium) for the various actions in the financial analytics domain and with existing related Commission projects.
3) Publicity: Non-confidential aspects of the project may also be published (both in print and online media) in appropriate “popular” scientific journals and magazines, trade journals, conferences, etc.
4) Project Workshop: Workshop for the targeted community will be held in UK or Brussels organised by EurActiv in conjunction with Handelsblatt in Month 34.
5) SSIX Training Wiki: Contains all training materials, planned training events, links to source code repositories, API’s and documentation.
WP66.8Technology Transfer and Dissemination Plan 2nd VersionThe Technology Transfer and Dissemination Plan - 2nd Version outlines our adopted dissemination and communication strategy. In addition, it provides a reflection on the use of social media, including a section on best practice for social media dissemination. Furthermore, a section dedicated to the stakeholder outreach of the communication activities implemented so to date. As a next step, the document focuses on the impact evaluation of various dissemination and communication tools like for example brochures, videos, newsletter, workshops, etc. Additionally, the dissemination activities of the SSIX Brexit case study are elaborated, including a special report, extra blog posts and increased social media activity.28/02/2017
WP66.9Technology Transfer and Dissemination Plan Final VersionThis deliverable has three main focus points. In the first part (Section 1) it focuses on the technology knowledge transfer model, which is aiming to transfer knowledge and skills generated in the course of the project. This transfer of expertise is achieved through a series of training sessions planned and implemented by the SSIX partners. The various formats considered most suitable for the training sessions vary depending on the complexity of the topic and on the level of confidentiality. Therefore it is differentiated between online and offline training sessions as well as between internal and external ones.
In the second part (Section 2) the deliverable focuses on the implementation of the dissemination activities implemented during the last phase of the project. Supported by several examples it provides an overview of all key communication measures that took place. One of the project highlights was the SSIX final event which is elaborated on in the third part (Section 3) of this document. Finally Section 4 concludes and summarises the deliverable.
WP77.5Pilots in IndustryPerforming ‘real-world commercial trials of the SSIX tools and services with industry partner customers’ is part of the validation of the technology and the methodological approaches developed in the SSIX project. The platform and the NLP pipeline provide tools and results for the live-analysis of social media streams on financial products, companies and goods as well as self-defined topics and named entities such as political parties or people. During the years 2016 and 2017, the commercial partners of SSIX demonstrated that they have been able to use the outcome of the research project, integrate it with their own business needs and environment and create value for their clients.
Both, the methodology and the analysis of using modern technologies like Machine Learning (ML), Entity Detection as well as Sentiment Analysis as input for specialised X-Scores for the financial domain as well as the concrete results of the SSIX platform. With its configurable live feed from social media, its capability of annotating raw data with categorizations, entity linking, and sentiment data has been used in the industrial pilots.
The industrial partners gained insights into new methods and technologies and they used the pilots to test new business opportunities with and for their clients. From their core business, they can now extend their products and services offering.
Feedback from end users and the SSIX industry partners has been used in further developing the capabilities of SSIX platform and stipulated new features and usages which have not been foreseen in the original approach.