Gain insights into conducting meaningful data analysis in Python without prior programming experience. Discover Python fundamentals, data preprocessing, EDA, and predictive analytics for solving business problems.

env.tar.gz

Python3

Ready to move past Excel for complex business analysis? Then you’ll find this course very helpful.

This hands-on introductory Data Science course is aimed at professionals and students who don't have any experience with programming. It will help you advance your career by preparing you to conduct meaningful data analysis in Python on any dataset — large or small.

You’ll begin with the fundamentals of Python, with focus on CSV files in Python, covering concepts like data preprocessing and Exploratory Data Analysis (EDA). In the second half, you'll focus on predictive and inferential analysis using statistical and machine learning techniques, and learn how these techniques can help solve business problems.

Data Science for Non-Programmers

# Ensembles
Recall that in the lesson [Random Forests](https://www.educative.io/courses/data-science-for-non-programmers/random-forests/), we learned that an **ensemble** is a collection of different predictive models that collectively decide the predicted output. Ensemble methods are divided into two categories:

* Bagging 
* Boosting

# Bagging 
In **bagging**, each individual model randomly samples from the training data with replacement. This means each model is different.

Note that we do not train individual models on random subsets of the data; rather, they are trained on the whole data set, but each training example is randomly sampled with replacement. For instance if our training data has 6 numbers such as [1,2,3,4,5,6] and we sample 6 times with replacement, we might get [1,2,2,4,5,5]. Therefore each individual model is different. 

In Bagging, the result is obtained by averaging the responses of the N models or majority vote.


# Boosting
In **boosting**, each individual model samples from the data, but this time, the sampling is not random, rather it is weighted. This means that in contrast to bagging, where each example in the training set has an equal chance of being chosen for training, in boosting, different samples have different chances of being chosen. 

Therefore, individual models are trained on different training sets, and hence, different models are obtained. 

In Boosting algorithms, each model is trained on data, considering the previous model’s success. After each training step, the weights are redistributed. Misclassified data increases its weights to emphasize the most difficult examples. In this way, subsequent models will focus on them during their training.

Once all of the models give their output, the second stage is to decide the final output. the final output is again based on weights. Boosting assigns a second set of weights, this time for the N models, in order to take a weighted average of their predictions. In the Boosting training stage, the algorithm allocates weights to each resulting model. A model with good classification results on the training data will be assigned a higher weight than a model that performs poorly. 

Two common boosting algorithms are:

* Gradient Boosting 
* AdaBoost

These differ in their loss functions. AdaBoost can be considered a special case of Gradient Boosting with *exponential loss*. Whereas Gradient Boosting is more flexible as it uses the entropy loss we used with logistic regression. We will use both of these later in this lesson.  


This lesson will focus on how to use boosting and bagging machine learning algorithms in Python.

What is Data Science

Python Basics

Handling Tabular Data in Python

Data Cleaning

Exploratory Data Analysis

Statistical Inference

Predictive Models

Machine Learning

Ensembles: Bagging vs Boosting

Ensembles

Bagging

Boosting