Blogs

19
Jan
An economist… working on the good side
  • Sergio Mayor
  • 740 Views
  • 0 Comment
  • Big Data . Big Data & Business . Big Data Consulting . Big Data Value . Carme Artigas . Fraude Fiscal . Synergic Partners .

Before I start, I would like to introduce myself to better contextualize this post. My name is Sergio Mayor Martin, I am an economist and I am currently working as a Business Consultant at Synergic Partners, pioneer in the world of Big Data in Spain. For us, the students of Economic Sciences, the name of our discipline has always been under discussion, with the second half of the title of those studies often being left off and leaving us with just plain “Economics.” This is due to traditional thinking, based on everything the term “Science” carries along with it, what exactly it should be and why economics is not one. This affirmation with reference to the world of economics is…

22
Nov

If I ask about the best way to manage the storeroom of a neighborhood shop, I am sure that we would quickly reach the conclusion that there are certain aspects that are absolutely necessary, such as, for example, that it be organized (in order to find the product when a customer asks for it), that it have an inventory (in order to know how many units I have, which models, their conditions and whether or not they are close to expiration) and that it be locked (in order to avoid someone coming and stealing the merchandise or simply destroying everything so that we cannot sell it). If, instead of the neighborhood shop, we were to discuss Amazon, no one would…

08
Nov
A Brief History of Machine Learning
  • Víctor Gonzalez
  • 1416 Views
  • 0 Comment
  • Big Data . ciencia de los datos . Data Science . Machine Learning .

As members of the Machine Learning community, it would be a good idea for us all to have an idea of the history of the sector we work in. Although we are currently living through an authentic boom in Machine Learning, this field has not always been so prolific, going through periods of high expectations and advances as well as “winters” of severe stagnation. Birth [1952 - 1956] 1950 — Alan Turing creates the “Turing Test” to determine whether or not a machine is truly intelligent. In order to pass the test, the machine must be capable of making a human believe that it is another human instead of a computer. 1952 — Arthur Samuel writes the first computer program…

18
Oct
IoT: When Material Acquires Intelligence
  • David_Morillas
  • 1593 Views
  • 0 Comment
  • No tags

The term IoT refers to the Internet of Things, a sort of interaction between the digital world and the physical one, a platform for connecting people, objects and environments. In the environment of IoT, the moment in which something connects, it becomes capable of accessing the power of computation of the cloud or, in other words, it becomes intelligent. The implementation of this technology and the trend towards automatization is unstoppable; each second an average of 127 things establishes a connection with the internet for the first time. All of this is possible thanks to what Moore’s law has not been able to accomplish, which says that as electronic computation grows in potential, it decreases in price exponentially. These improvements…

17
Sep
Warnings about normalizing data
  • Santiago Morante
  • 1607 Views
  • 0 Comment
  • No tags

For many machine learning algorithms, normalizing the data of analysis is a must. A supervised example would be neural networks. It is known that normalizing the input data to the networks improve the results. If you don't believe me it's ok (no offense taken), but you may prefer to believe Yann Le Cunn (Director of AI Research in Facebook and founding father of convolutional networks) by checking section 4.3 of this paper. You can catch up the idea with the first sentence of the section: Convergence [of backprop] is usually faster if the average of each input variable over the training set is close to zero. Among other things, one reason is that when the neural network tries to correct…

15
Jun
When Michelangelo Invented Big Data
  • Álvaro Alegría
  • 2077 Views
  • 0 Comment
  • Big Data . digital transformation .

They say that when Michelangelo was asked about his impressive technique in sculpting the Pietà in just one piece, his answer was: “The sculpture was already inside of the stone. I merely eliminated the extra marble.”   It’s impossible to deny that, from a simplistic point of view, sculpting is getting rid of the extra material. However, something must influence the artist, because, if most of us were handed a block of marble, a hammer and a chisel, we would most likely not be able to sculpt anything recognizable.   And it is the creativity of the artist, more than the handling of their tools, that is the added value that they bring to what they do and what makes…

08
Apr
An architecture to tweet them all
  • Jesús Sánchez
  • 2424 Views
  • 2 Comments
  • No tags

The definition of the project was simple: find out what it is that people think of our client on Twitter, both the good and the bad, and be able to visualize it all on a dashboard. Or at least as much as possible. As soon as we find out the range, we try to get as close as we can. Read from Twitter, analyze what is said and process it. Three modules. The architecture didn’t seem complicated, but we had no idea what was ahead. I have always been taught that a good engineer finds a problem, suggests a solution and chooses a series of tools to carry it out. Problem, solution and tools. Always. Well… not always. In this…

17
Mar
Big Data, the Key to Business Value
  • Albert Solana
  • 2427 Views
  • 0 Comment
  • Big Data . Business . conectividad . datos . innovación . internet of things . monetización . negocio . Value of Data .

We live in a time in which digital transformation is a part of every forum on business. In these forums, an idea is spreading that the digital will allow brands to interact directly with consumers, to perceive the perception of the brand, and to understand how their products are being used. Therefore, the digital environment will allow them to better understand the consumer and will serve as an anchor for customer loyalty in both the medium term and the long term. Companies have already realized how disruptive this new context has become, and the large investments made to have a presence in digital channels can be clearly seen, whether it be in owned media (space belonging to the brand itself),…

29
Feb
The hottest trends in Big Data and Data Science
  • Synergic Partners
  • 3439 Views
  • 0 Comment
  • Big Data . Big Data Diet . Data . Data outsourcing . data plumbing . Data Science . Deep Learning . Hadoop . HPC . light analytics . small data . Transformación digital .

Based on client surveys – vendors of products and data processing platforms – as well as trends on popular blogs, LinkedIn and similar posts, here are the most in-demand big data and data science topics: The importance of data plumbing (as the cleaning and preparation of data has come to be known in general terms) in optimizing big data tools, making them more precise, safer, more reliable and faster through “data pipes” (internet, intranet, in-memory, local servers, the cloud, Hadoop clusters, etc.), optimizing such aspects as redundancy, load balance, and the intake, storage, compression and summary of data, among other things. The rise of the data plumber, an architect of systems and systems analysis (a new figure in the ranks…

28
Jan
XGBoost
  • Roberto Izquierdo
  • 4109 Views
  • 0 Comment
  • Kaggle . Machine Learning . XGBoost .

In Kaggle machine learning competitions, two techniques tend to be dominant: the use of groupings of decision trees for structured data and neural networks when the data includes images or sound. Traditionally, Random Forest dominated competitions in structured data, but another algorithm has recently surpassed it in these competitions: Gradient Boosted Trees. Like RF, GBT classifies examples through the use of a grouping of decision trees. In the case of the latter, the trees are constructed sequentially, adding at each iteration the tree that best compensates for the errors in the other trees. This method is called gradient because the model evolves tree by tree towards a minimum of error. The tool used in these cases is called XGBoost, an…

Featured customer