Call Us: +302103805826 - Mail: info@paloservices.com
PaloAnalytics

PALO Ltd

logo-paloservices

PALO Ltd. was established in 2008 and currently operates in 5 countries: Greece, Cyprus, Serbia, Romania and Turkey. In the first 3 countries it has a distinct corporate presence while in Romania and Turkey the service and support is provided by Greece. PALO Ltd now has 3 services / products in 5 countries. Its services are palo web sites palo.gr, palo.com.cy, palo.rs, palo.ro, palo.com.tr, palo pro and the mobile apps Palo News Digest.

Palo web sites (gr, rs, ro, com.tr, com.cy)

Palo is the largest search engine for news in Southeast Europe. They are websites that collect news from many different sources and centralize them in a user-friendly way for the end user. The basic processes performed in the kernel of palo are:

Crawling

In this process, the mechanism having a list of sources collects information (news by interval). This process is continuous and depending on the type of source it becomes more or less sparse. On a general level (palo.gr & palo pro) this mechanism is responsible for the collection of data from heterogeneous sources such as multimedia pages, blogs, social networking pages such as facebook, twitter and others. The crawler works as a data collection service and also contains internal mechanisms for self-improvement and scalability. It works continuously and collects additional metrics that help to better monitor it, but also to function better.

Classification

A procedure according to which one or more categories are assigned to each item entered into the system. The categories are predefined and the assignment is in accordance with the association with some or some categories. It operates continuously and regularly, self-training for the ultimate classification of the data. Finally, this mechanism also collects several metrics that are used both internally and externally by surveillance systems.

Clustering

Clustering is the process of grouping news according to their similarity. Due to the many sources and the natural process of news reproduction from many different media, the system groups the same / identical news and presents them as one. This mechanism is the central nucleus of palo. It focuses on data from articles rather than social networks and its purpose is to update existing and new clusters with related reports. Its mode of operation is quite complex and differs from classical clustering techniques due to the nature of the data. Metrics and its parameterization variables are numerous and constantly changing for better results.

Palo Hot Topics

This mechanism deals with the extraction of specific words or phrases reproduced from all reports at regular intervals capturing the current news reporting of the space mentioned.

Palo Related Clusters

The mechanism is responsible for associating clusters based on internal criteria for better cluster association not only on content, but also on category and time.

Summarization

Abstract extraction is the process by which the system receives as input the body from the set of cluster articles and extracts a brief and comprehensive summary of all the texts. This algorithm has the ability to select suggestions that give the key points the reader must know from a cluster of news on the same topic. This software reads all texts from a cluster of news and synthesizes suggestions that give the summary in the form of bullet points. The size of the proposals and their number is customizable and can be adapted to our needs.

PaloPro

PaloPro is a separate subscription service / platform. In addition to data from news sites and blogs, reports from social media are also collected and some additional procedures are being made:

Crawling

This process gathers data from sources such as:

  1. facebook
  2. twitter
  3. Ιnstagram
  4. youtube
  5. forums
  6. News sites
  7. Blogs

Entity Extraction

From all the documents that are collected, the entity is originally extracted. Exporting entities is the process by which the basic terms referring to a text can be extracted, which can be of many kinds and relate to: person, company, product, etc.

Sentiment Analysis

For every data entered into the system there is sense information about the entities it contains. More specifically, there is information about positive, negative and neutral mood for every data.

Platform’s Functions

The basic function of the system is to create a workspace where the user completes the keywords and entities that interest him to follow. The mechanism then finds the data containing the requested keywords and / or entities and gathers them on a dashboard page. Beyond general statistics and graphs the user has a deeper analysis using widgets. The most important widgets are:

  • Top Influencers: It contains a list of top influencers, ie the sources / users that most influence the brand or product we are watching. It is information derived from a metric per data stream.
  • Top Referrers: Contains a list of users / sources with the most reports for that workspace.
  • Topics: Contains a list of clusters that have articles for that workspace.
  • Sentiment Radar (24h): It is a graph that gives the polarity of the last 24h.
  • Top Quotes: Contains a list of the most important quotes that have been found in articles.

Research Interests Palo

Palo, as an innovative company, has set as a priority its active participation in research fields with the aim of developing both existing services and creating new ones that will change the information market. Palo’s main research fields relate to large data recovery and analysis (Big Data). The data gathered come from sources such as news websites, blogs, forums, etc., as well as from social networks (facebook, twitter, etc.). These data belong to 2 categories: texts and multimedia. The challenges posed when retrieving and analyzing text data are many. Palo wishes to continuously improve text data management processes through research in areas such as:

  • Smart ways to automatically retrieve text from the web. Internet sites are “alive” and constantly changing, which makes it impossible for them to be monitored by humans and requires automation. This field is one of the most important, as it is the source of the data.
  • Classification of extracts (or whole) of clustering. This process allows you to view and explore texts with similar content.
  • Edit the text data in a way that gives the text a semantic character. For example, finding nominative entities, categorizing text, mining views from a text, finding a feeling depicted in a text, etc.
  • Discover networks / knots of influence. As the information diffuses on the internet and on social media, it is important to be able to export information such as:
    • From where the original information came from. Which is the point that first introduced the information into the network.
    • Which are the nodes that most influence the diffusion of information.
    • Who are the nodes that reproduce the information itself or transform it.
  • Find reports that have a sharp and rapid diffusion. Such reports are, for example, news stating an important event.
  • Ways in which references are mutated over time. For example, the course of a news release from the time it was first published until its release and expulsion.
  • Discover reports that present false information. This is particularly important as false information can influence public opinion.
  • In addition to text data, media is an important piece of information from which valuable information can be extracted. Useful results for this purpose can offer the following fields:
    • Discovering a correlation between a text and the image that accompanies it. In this way, images suitable for a text can be selected.
    • Categorize images to avoid showing images that are inappropriate or offensive.

Relevant Projects – Distinctions

Additionally, the company has previously participated in funded activities. More specifically, in 2015 the project was completed with a project code: ICT -000651 (partnership with S. CHARITAKIS AND CO.) Under the “OP DIGITAL CONCISE”, with a budget of 750,000 euros. Additionally, it has implemented a project under the “New Innovative Entrepreneurship” OP Competitiveness, with a budget of € 68,722. In addition, the company has submitted a proposal under Horizon 2020 (H2020-SMEInst-2014-2015) for the Open Disruptive Innovation Scheme, for which it has received a seal of excellence as an innovative proposal for implementing the PALO AGORA service .

Finally, it should be noted that the company has been distinguished and received the distinctions: Innovation and Extroversion in PAPASTRATOS Start-Up / Scale-Up Greece Awards 2017 Silver Award at “Top Innovation in Exports” of “Greek Export Awards 2016” the Palo News Digest). Silver Award at “Mobile Excellence Awards 2016” in “Innovative apps” category (for Palo News Digest) Merit Award at “Greek Graphic Design & Illustration Awards” EBGE 2016 (for Palo News Digest) Palo Pro was awarded at the “E -volution Awards 2014 “as a dynamic new entrant. “E-volution Awards” 2013 (in the Research and Development category for palo.gr). Ermis Awards 2012 (in the media category for palo.gr).

Back to Top