Neural Networks for Image classification: Tensorflow and Keras

1. Introduction One of the main interests of AI nowadays is computer vision. This field includes several different tasks, like image classification (i.e. classify an image as a dog or a cat), object detection (locate where exactly the dog/cat is), image segmentation (which are the pixels that belong to the object detected) and others. The applications are countless, from self-driving cars, number-plate recognition or face recognition systems to medical applications such as cancer detection and many, many more.

Regex is all you need? Introduction to Parsy

Introducción Many data scientists have found themself in the situation of extracting useful information from raw text: parsing a date, a telephone number or a URL. The most known and used tool for such a task are regular expressions. They are powerful and are supported by any programming language we pick, but they suffer 3 problems: Long term Maintainability. As an example, the following regex parses a date in format yyyy-MM-dd HH:MM:SS with optional dashes and colons: r'^(\d{4})-?

Air Pollution

Introduction The air pollution and the climate change have an effect on our lives that nowadays is directly noticeable and increases a need to know the air quality of our city. Autonomous Communities and City Councils offer open access of air quality data in real time as well as historical data. We’re going to see how to access the air pulltion data from several cities in Spain and how to analyze these data.

How to connect PostgreSQL with Python

Introduction The pandas.io.sql is such a powerful and useful module to interact with relational databases seamlessly. However, the read_sql and to_sql functions do not solve every posible use case. Particularly, when it comes to writing to sql, the latest pandas version (1.0.3) does not have implemented an update method if the data to be inserted alrealy exists. Moreover, when working on complex solutions with frequent interactions with the database having an abstraction layer that deals with the connections, building of queries and so on should be a must.

How to draw maps in R using sf and ggplot2

Introduction The use of maps to represent geographic information provides very valuable context for a better understanding of the data. Maps are relatively easy to comprehend for most people, which allows key insights to reach a broader audience. However, working with maps is not easy. A map represents a projection in a 2D plane of a territory that’s actually on the surface of a 3D sphere. This implies a considerable amount of technical challenges and solving them is not a trivial task.