Feature engineering: secret weapon for improving Machine Learning Solutions!!!

Feature engineering: secret weapon for improving Machine Learning (ML) Solutions!!!

These days we have developed some examples of Geospatial Maps with CDMX Open Data, in particular historical data from Linea Mujeres, until Nov 2020 that was migrated to LOCATEL

We generated a heat map, with GeoPandas and Folium, of CDMX that puts colors according to the number of calls from women

https://catedraunescodh.unam.mx/catedra/Papiit2017/mapas/MapaLineaMujeresDatosHistoricos202011.html

The data was obtained from:

https://datos.cdmx.gob.mx/dataset/linea-mujeres

Another that offers access to data in Federal Entities

https://catedraunescodh.unam.mx/catedra/Papiit2017/mapas/MapaCUDH_UNAM_CEDAW_DerechosPoliticosMujeres_MarcoNormativoEstatal.html

They are examples of how to display information with GeoSpatial Maps using Python, Folium and soon with Streamlit and inserting Time Series and numerical prediction with Prophet

Only open source libraries and open data with Python, don't miss our next workshops on these topics and also on NLP for text handling with SpaCy!!

We have already recorded several tutorials on the subject.

They are the shorts of the film (Trailer) before the premiere in 2022.

Free access on YouTube:

Conference: Data Science with a Gender Perspective. LOCATEL case

https://youtu.be/ww8wdOniWUw

A first tutorial in Python with Jupyter Notebook

https://youtu.be/Lw__tdbgRlI

Second tutorial

https://youtu.be/kWJmvJWOJq4

Third tutorial

https://youtu.be/u7dI0CJrn6A

Forth tutorial

https://youtu.be/mJCsnVO-nmA

Fifth tutorial

https://youtu.be/ayYXY18Jbac

Sixth tutorial

https://youtu.be/cDIjkZlWACM

Wednesday, March 02, 2022 at 7:00 p.m.

Course with RECOVERY FEE TO RECEIVE THE MATERIAL: 900 pesos

Workshop duration: 2 hrs

It is the Quality Ingredients Process!!!!

Great models cannot exist without great quality data.

Now we say: “machine learning is basically feature engineering”

A feature as an attribute/column of data that is meaningful to an ML model.

Now Datasets have a large number of columns compared to the number of observations.

This can lead to what is known as the curse of dimensionality which describes an extremely sparse universe of data that ML models have difficulty learning from Interpretability of the data and model is key.

Feature engineering is the most interesting technique now!!

Feature engineering is the methodology used to extract numerical representations from unstructured data for an unsupervised (trying to extract structure from a previously unstructured dataset) model.

Feature engineering is transforming data into a format that optimally represents the underlying problem that an ML algorithm is trying to model

How we use algorithms and statistical testing procedures to identify the strongest features?

Stop tuning your ML model; learn more about features engineering; it will change your life.

Deliver huge improvements to your machine learning pipelines without spending hours fine-tuning parameters!

Feature engineering is the secret weapon for improving your machine learning’s output.

By enhancing the data ingestion, manipulation, and transformation elements of your pipeline, you can see dramatic improvements in your downstream results without endlessly fine-tuning parameters or chasing the latest models.

Without understanding the data, it is impossible to capture, learn from, and scale up the patterns locked within the data.

The CRISP-DM is a Machine Learning Pipeline in 6 steps. The first 3 are strategic and close related to Feature engineering.

We have our practical methodology of 12 tasks that we share at the workshop for the followings aspects of our CRISP-WebGestiones Methodology

1. Defining the Problem Domain

2. Obtain data and Exploratory Data Analysis

3. Feature Engineering. Prepare data as numerical features for the algorithms

On Friday, December 10, 2021, we closed our 2021 Data Science cycle.

Course with gender perspective: Python, Matplotlib, Seaborn, Folium, GeoPandas and Streamlit deployment with Open Data from CDMX associated with women

In the workshop we use Open Data from CDMX to show how we carry out the Exploratory Data Analysis (EDA) to determine feature importance with Feature engineering methods.

Much of the current discourse around Artificial Intelligence (AI) and Machine learning (ML) is inherently model-centric, focusing on the latest advancements in ML and deep learning.

This model-first approach often comes with at best little regard and at worst, total disregard to the data being used to train said models.

Workshop duration: 2 hrs

RECOVERY FEE TO RECEIVE THE MATERIAL: 900 pesos

Check our YouTube channel

https://www.youtube.com/channel/UCf49zOpIRJo5zK9P0Ma-sFQ/videos

Paypal payment

PayPal.me/saxsa2000

Bank payment, request CLABE BBVA

Dr Gabriel Guerrero

saxsa2000 (at) gmail.com