Data Science & Machine Learning(Theory+Projects)A-Z 90 HOURS

Course content:

1. Python for Data Science and Data Analysis

a. You start with problem-solving and finish with fancy indexing and plots in Matplotlib.

b. No prior knowledge in any computer science language is assumed.

c. Great fun with Python language.

d. Reasonable treatment of data science packages (NumPy, Pandas, Matplotlib, Seaborn, and Sklearn).

e. After this course, you will be a competent Python programmer as well as a reasonable expert of data science packages (NumPy, Pandas, Matplotlib).

f. This section is designed to teach you programming in general also. Therefore, shifting from this language to any other language after this section is not difficult.

2. Data Understanding and Data Visualization with Python

a. This section deals with the in-depth treatment of data science packages both for data manipulation as well as data visualization.

b. While Section 1 focuses more on Python language, this section focuses completely on data science packages and their efficient use.

c. The packages covered in this section include NumPy, Pandas, Matplotlib, Seaborn, Bokeh, Plotly, and Folium.

d. As far as we know, this is the most comprehensive section on data understanding and visualization among the available ones.

e. Further, this section is designed to reduce the dependency on core Python language to be treated independently, as well.

f. 2D and 3D visualizations, interactive visualizations, and geographic maps are also covered.

g. Proceeding in data science with being able to effectively play with the data using famous packages makes progress much worse, and this section addresses this concern.

3. Mastering Probability and Statistics in Python

a. Obviously, concepts in data science are not new. In fact, it is also believed that data science is merely a renamed version of Probability and Statistics. Well, without being biased to that extent, we will say that the practical nature of applications was uncovered earlier even though the theory traces back to Probability and Statistics.

b. One way or the other, knowing Probability and Statistics makes a significant theoretical as well as practical difference.

c. Most of the courses on Probability and Statistics, however, fail to link the data science practices and theory by merely focusing on the axiomatic treatment of the subject.

d. We build this section by keeping the practical needs of data science in mind as well as the importance of theory.

e. Wherever important, we deliberately explain and show the relationships by derivations and even through Python Code.

f. This section builds a very sound basis for understanding the classical concepts in data science as well as its more recent generalizations.

g. We start with the very basics of Probability, go through inference and estimations, link famous machine learning techniques with conditional probability, and finally, show that Deep Neural Networks indeed learn a probability function eventually.

4. Machine Learning Crash Course

a. Although several concepts, or even all, fall under the umbrella of Probability and Statistics, it turns out that most of the concepts have made their own practical place, mostly derived through engineering, with the name of Machine Learning. For example, the term “overfitting” is now referring to the area of machine learning.

b. Machine Learning brings its own set of practices to reach the demands of automation. Hence, mastering these concepts becomes inevitable.

c. This section is actually a quick walkthrough of the concepts in Machine Learning and focuses on all the theoretical as well as practical concepts.

d. We mostly cover applications using the Sklearn Python package and build machine learning pipelines in this section.

e. We also elaborate on more advanced areas of machine learning, which we later present as separate sections.

5. Feature Engineering and Dimensionality Reduction with Python

a. Knowing the sections you have covered thus far certainly brings you a huge clarity of the field. But there is still one thing that brings the improvements in the results with a reasonable margin, and that is data preprocessing or data preparation.

b. Most of the data science today relies on preparing the data suitable for machine learning models. An effective way of data preparation, most of the time, becomes a game-changer.

c. This section focuses on data preparation for machine learning models.

d. We build this section to provide an understanding of why selecting features and transforming features are important.

e. We also discuss practical issues with real data, like missing values and non-numeric data.

f. We discuss the performance improvements both in terms of execution time as well as the accuracy of the models.

g. We explain the required mathematical background in a simple way.

h. Finally, all the concepts are made more easily understandable by coding relevant examples in Python.

6. Artificial Neural Networks with Python

a. With the availability of a huge quantity of data as well as computation power, a relatively old machine learning model, Artificial Neural Network turns out to be the game-changer in data science.

b. Artificial Neural Network can approximate almost any pattern in the data. Further, it has a much greater data utilization capacity as compared to the more classical methods.

c. With the recent rise of ANNs, a lot of practical techniques are also discovered, particularly for ANNs.

d. Also, working with a large amount of data brings its own challenges for learning algorithms.

e. In this section, we address all these concerns and cover ANNs in depth.

f. We also introduce another framework, “TensorFlow,” for working in ANNs.

g. With this section in hand, you can now target much larger machine learning problems.

7. Convolutional Neural Networks with Python

a. ANNs, in its most basic form, is not that suitable for image data and for the problems in computer vision.

b. Convolutional Neural Networks (CNNs) are considered a game-changer in the field of computer vision. CNNs are not limited to images only. You’ll find them everywhere now, from audio processing to more advanced reinforcement learning (i.e., Resnets in AlphaZero). So, the understanding of CNNs becomes inevitable in all the fields of data science. Even most of the Recurrent Neural Networks (RNNs) rely on CNNs nowadays.

c. In this section, you will to learn about:

i. The significance of CNNs in data science.

ii. The reasons to shift to CNNs from hand engineering (classical computer vision).

iii. The major concepts from the absolute beginning with complete unfolding with examples in Python.

iv. Practical explanation and live coding with Python.

v. Evolution of CNNs — LeNet (1990s) to MobileNets (2020s).

vi. Intricate details of CNNs including examples of training CNNs.

vii. TensorFlow (Google’s deep learning framework).

viii. The use and applications of CNNs (with implementations in framework TensorFlow) that are more recent and advanced in terms of accuracy and efficiency.

ix. The use and applications of pre-trained CNNs (with implementations in framework TensorFlow) for transfer learning on your own dataset.

x. Building your own applications for Human Face-Verification and Neural Style Transfer.

After completing this course successfully, you will be able to:

• · Relate the concepts, principles, and theories in Data Science & Machine Learning.
• · Understand the methodology of Data Science & Machine Learning using real datasets.

Who this course is for:

• · People who want to become perfect in their data speak.
• · People who want to learn Data Science & Machine Learning with real datasets in Data Science.
• · People from a non-engineering background who want to enter the Data Science field.
• · People who want to enter the Machine Learning field.
• · Individuals who are passionate about numbers and programming.
• · People who want to learn Data Science & Machine Learning along with its implementation in realistic projects.
• · Data Scientists.
• · Business Analysts.