90Days Data Science Bootcamp: Build Portfolio Of 90 Projects

Explore Python basics, its uses in data science, analytics, machine learning, and web development, and core data types like integers, floats, and strings using Google Colaboratory.

Explore variables in python as containers that hold data, learning how integers and strings are stored in memory, referenced by names, printed, and inspected with type.

Explore lists in Python, a versatile data type that can hold integers, strings, and characters with zero-based indexing. See how lists are mutable and update items by index, unlike tuples.

Explore Python lists from creating empty lists and appending items to inserting at a position, copying behavior, and list operations like pop, reverse, sort, extend versus append, and split.

Discover how dictionaries in python store data as key-value pairs, access items with dict.items, and update or extend values—such as lists—using append, keys, and values.

Explore Python tuples by understanding their syntax, immutability, and how to access elements through indexing and slicing, while converting tuples to lists to compare types.

Learn how sets in Python enforce uniqueness, cannot be indexed, and differ from lists; initialize a set, add elements, and print to see distinct items.

Explains Python if else statements, condition evaluation, and indentation for blocks, with examples like checking if a number is greater than nine and handling unrecognized names.

Explore Python for and while loops, including iterables, ranges, and dictionary iteration. Learn unpacking items, values, and building lists with list comprehension and squaring numbers.

Harness Python list comprehension to produce a new list by applying an operation to the original list, such as squaring even numbers, resulting in concise, readable code.

Explore how while loops in Python operate: evaluate a condition, print values with range, and use break, noting that the else block runs only when the loop ends naturally.

Learn how to define and call functions in Python, pass parameters, and reuse code; compare using print versus return and how booleans indicate even or odd.

Learn bankcard tampering detection with computer vision to verify IDs like Aadhaar and Pan cards, using ten steps to upload images and compute a similarity score.

Use Google Colab to run a pan card tampering detection project, calculating a structural similarity score between the original and tampered images downloaded from URLs.

Build a pan card detector with OpenCV by resizing images, converting to grayscale, and comparing original and tampered images using structural similarity, thresholding, and contours.

Create a Flask app for a model, with a project structure (config.py, app.py, __init__, views.py), development and debug configs, and file upload with image processing tools.

Create an image classification web app by uploading images, processing with cv2 (resize, grayscale, contours, bounding boxes), and rendering results with a materialize css front end.

Deploy a Python app to Heroku by creating a virtual environment and installing requirements. Upload code to GitHub, connect GitHub to Heroku, and deploy main branch to obtain a URL.

Test the deployed pan card detector by uploading a file and checking tampering with 28.57% correct; deploy via GitHub and Heroku with a front end.

Explore dog breed prediction with a convolutional neural network built in Keras and TensorFlow, training on Kaggle data in Colab, including one-hot encoding, normalization, model training, evaluation, and predictions.

Train a convolutional network with Keras and TensorFlow in Google Colab to identify dog breeds, using Kaggle datasets and Kaggle API setup.

Load libraries and labels, inspect class counts, select three dog breeds for classification, one-hot encode targets, convert images to numpy arrays, and normalize pixels by 255.

Build a cnn sequential model with conv2d, max pooling, and flatten layers, then dense layers for a three-class classification using softmax activation with categorical cross entropy and Adam optimization.

Plot training vs validation accuracy across epochs using matplotlib, inspect model history, and assess a test accuracy around 83.33%. Explore hyperparameter tuning—layers, kernels, neurons, activation, optimizers, and initializers.

Learn to build a Streamlit ml app that loads a dog breed model, accepts image uploads, and shows predictions through a simple main app script.

learn to set up a data science app from the command line using Anaconda prompt, create a requirements file, install Streamlit, and run a local app to predict dog breeds.

Learn to add logo and text watermarks to images using OpenCV. Load images from URL, apply region of interest, and render watermarked output in Google Colab.

The lecture demonstrates using Google Colab and OpenCV to add logo and text watermarks to images for copyright protection, importing cv2, numpy, requests, and PIL and reading image URLs.

Learn to add watermarks to images by converting them to rgb numpy arrays and finding the center region of interest. Apply a logo or text with cv2.

This lecture builds a flask app for an image watermarking project, covering config, app setup, image upload views, and logo or text watermarking with Pil, numpy, and OpenCV.

Deploy the image watermarking project to Heroku by linking a GitHub repo, enabling automatic deploys from the main branch, installing Python dependencies, and testing the logo watermark and text watermark.

Build a multi-class traffic sign classifier with a CNN using TensorFlow and Keras, from Kaggle data in Colab to preprocessing, one hot encoding, training, and evaluation for self-driving cars.

Develop a traffic sign classification model with a convolutional neural network in Google Colab, using TensorFlow and Keras, after connecting Kaggle and downloading the German traffic sign dataset.

visualize and resize traffic sign images to a uniform 50 by 50 shape, divide by 255 to normalize, convert to numpy arrays, and apply one-hot encoding for 80/20 train–test split.

Design a convolutional network with conv2d layers, max pooling, and dropout, ending in a 43-neuron softmax classifier for multi-class classification, and tune hyperparameters with sparse categorical cross entropy using Adam.

Detail the test set preparation for traffic sign recognition, including cleaning files, scaling images to 50 by 50, normalizing, predicting, and reporting 43 classes with 95%+ accuracy.

Learn to extract text from images with Tesseract and Pytesseract, enhance results via OpenCV steps like blur, threshold, and erode, and draw rectangles around words to automate OCR.

Utilize Google Colab to run text extraction on online images with Tesseract, pytesseract, and OpenCV. Install Tesseract, fetch an image from the internet, preprocess, visualize, and extract text.

Extract text from images with pytesseract using custom configurations, then remove unwanted symbols and apply OpenCV preprocessing, including grayscale conversion, noise reduction, and thresholding.

Learn erosion and morphological transformations with OpenCV, perform canny edge detection and skew correction, and use template matching and pytesseract to bound words with rectangles and build a Flask app.

Create a flask-based optical character recognition app from config and routing to upload images, extract text with tesseract, and display results in a materialize and jquery front end.

Install dependencies from requirements.txt with pip, run the app with Python app.py, then test the optical character recognition extractor by uploading an image.

Learn to build a convolutional neural network for plant disease detection using TensorFlow and Keras, with data preparation, training, evaluation, and predicting test images for a farmer-oriented app.

Set up google colab, mount drive, and import essential libraries like numpy, pandas, matplotlib, and opencv to build and train a cnn with tensorflow and keras for plant disease classification.

Visualize the plant disease image dataset, convert images to a numpy array, balance three classes, split into training and testing, normalize, and apply one-hot encoding with Keras.

Build a cnn model plant disease classification using conv2d, max pooling, flatten, and dense layers with softmax for three classes; train with categorical cross entropy and adam, save and visualize.

Create a Streamlit plant disease detection app that loads a Keras model, uses class names like tomato bacterial spot and common rust, and predicts disease from uploaded leaf images.

Detect and count vehicles in images and videos with OpenCV and Haar cascades, using grayscale images, contours, and XML classifiers for cars and buses.

Detect and count vehicles in images and videos using OpenCV and Haar cascade classifiers in Google Colab, loading cars.xml and bus_front.xml assets and processing with Python libraries.

Transform images through grayscale conversion, Gaussian blur, and dilation, then use Haar cascades for car and bus detection, drawing rectangles and exporting the processed video output.

Build a Flask vehicle detection app that uses OpenCV cascade classifiers to detect and count cars from uploaded images, then deploy via GitHub and Heroku.

Master face swapping using OpenCV and Dlib by extracting facial landmarks, triangulating, and performing a seamless clone to blend a source face into a destination image.

Learn to perform face swap in Google Colab using OpenCV and Dlib to extract 68 face landmarks from source image and swap with destination image using a pre-trained model.

Convert images to numpy arrays and grayscale, detect facial landmarks with dlib, perform triangulation, warp triangles, and perform a seamless clone to swap the source face into the destination image.

Create a flask face-swap app by outlining config, building the app and views, and handling image uploads. Explore front-end setup with index.html and materialize css for the user interface.