E-Commerce Shipping Prediction
- Built an End-to-End Web App using Data Science & Machine Learning for an
International E-commerce Company to predict whether their products will
reach on committed Delivery Time or not.
-
Collected the dataset from Kaggle.
-
Preprocessed the data well and built an ML model by tuning the Hyperparameters of RandomForest Classifier using RandomizedSearchCV.
-
Saved the Best Performing model in a Pickle file.
-
Built a Web App using Python at the Backend with Flask API and HTML & Bootstrap serving at the Frontend.
-
Deployed the Web App on Heroku Cloud.
-
The Web App receives the data from the User Input, makes predictions with the saved ML model and sends a Prediction text indicating whether the order will reach on time or not.
- Using such models, can increase the customer retention and customer
satisfaction.
Harvestify
- A simple ML and DL based website which recommends the best crop to grow, fertilizers to use and the diseases caught by your crops.
-
Farming is one of the major sectors that influences a country’s economic growth.
-
In country like India, majority of the population is dependent on agriculture for their livelihood. Many new technologies, such as Machine Learning and Deep Learning, are being implemented into agriculture so that it is easier for farmers to grow and maximize their yield.
-
The following applications are implemented:Crop recommendation, Fertilizer recommendation and Plant disease prediction.
-
In the crop recommendation application, the user can provide the soil data from their side and the application will predict which crop should the user grow.
-
For the fertilizer recommendation application, the user can input the soil data and the type of crop they are growing, and the application will predict what the soil lacks or has excess of and will recommend improvements
-
Built a Web App using Python at the Backend with Flask API and HTML & Bootstrap serving at the Frontend.
-
Deployed the Web App on Heroku Cloud.
- Using such website, can help the farmers in increasing their crop yeilds and hence contributing to the economy of the country.
Diabetes Prediction Web App
- Objective is to predict whether the patient has diabetes or
not based on various features like Glucose level, Insulin, Age, BMI and
various other features.
- The web app is developed to help people determining whether, they can be
diabetic or not in early stages. Early detection of diabetes can prevent the
adverse effects of it.
- Deployed the app using streamlit.
Black Friday Sales Prediction
- Predicting the amount spend by the customers during the sale of Black Friday
following all the steps of a data science lifecycle.
- Also built a web app using the Flask micro-framework and deployed using
Heroku.
- Predicting Customer behavior, can impact the decisions taken for the growth
of any organization.
- Using such models, can increase the customer retention and customer
satisfaction.
Credit Card Fraud Detection
- The problem statement chosen for this project was to predict fraudulent
credit card transactions with the help of machine learning models.
- The dataset is taken from the Kaggle Website website and it has a total of
2,84,807 transactions, out of which 492 are fraudulent.
-
The challenge in this project was that, it is highly imbalanced with only
0.17% transactions to be fraudulent.
-
The trained model got an accuracy of about 95%.
-
Aim is to help the credit card companies to make them capable to recognize
fraudulent credit card transactions so, that customers are not charged for
items that they did not purchase.
Census Income Prediction
- Predicting whether a person's income is above 50k or below 50k using various
features like age, education, and occupation.
- Performed all the steps from Data collection to Model deployment. Then
created a web app using Flask which is a python micro framework.
- Built various trained models using logistic regression, knn classifier,
support vector classifier, decision tree classifier, random forest
classifier and xgboost classifier.
- A hyperparameter tuned random forest classifier gives the highest accuracy
score of 92.77 % and f1 score of 93.08.
- Building such predictive models can help us better understand the population
of a country as well as the various factors affecting the economy.