Python Case Studies

Machine Learning business case studies solved using python. These are examples of how you can solve similar use cases for your own project and deploy the models into production.

I have discussed below points in each of the case studies.

1. How to explore the given data?
2. How to perform data pre-processing (missing values, outliers, transformations, etc.)
3. How to create new columns based on existing columns (Feature Engineering)?
4. How to select the best columns for machine learning (Feature Selection)?
5. How to find the best ML algorithm for the given data?
6. How to tune the predictive models.
7. How to deploy predictive models into production?
8. What happens after the model deployment?

NLP Use Cases

• TF-IDF Text classification
  Support Ticket Classification using TF-IDF Vectorization
• Sentiment Analysis using BERT
  Finding the sentiment of Indigo flight passengers based on their tweets
• Transfer Learning using GloVe
  Microsoft Ticket classification using GloVe
• Text classification using Word2Vec
  How to create classification models on text data using word2vec encodings
  

Regression Use Cases

• Zomato restaurant rating
  How to predict the future rating of a restaurant based on an ML model. A Case study in python.
• Predicting diamond prices
  Creating an ML model to predict the apt price of a given diamond.
• Evaluating old car price
  Predicting the right price for an old car using python machine learning.
• Bike rental demand prediction
  Create an ML model to forecast the demand of rental bikes every hour of the day.
• Computer price prediction
  Estimating the price of a computer, based on its specs.
• Concrete strength prediction
  How strong will this concrete be? Predicting the strength of concrete based on its mixture details.
• Boston housing price prediction
  House price prediction case study on the famous Boston data.

Classification Use Cases

• Loan Classification
  A predictive model to approve/reject a new loan application.
• German Credit Risk
  Classification of a loan as a potential risk or safe for the bank.
• Salary Band Classification
  Identify if you deserve a salary more than $50,000 or not.
• Titanic survival
  A case study to see what type of passengers survived the titanic crash.
• Advertisement Click Prediction
  A case study to predict if a user will click on advertisements or not

Deep Learning Use Cases

• ANN-Regression
  Creating an Artificial Neural Network model for Regression
• ANN-Classification
  Creating an Artificial Neural Network Model for Classification
• LSTM
  Predicting Infosys stock price using Long Short Term Memory network
• CNN
  Creating a face recognition model using the Convolution Neural Network