Data Science Portfolio

Here are some of my best Data Science Projects. I have explored various machine-learning algorithms for different datasets. Feel free to contanct me to learn more about my experience working with these projects.

Examining the effect of environmental factors and weather on Bike rentals

Skills Used: Python, Numpy, Pandas, Matplotlib, Seaborn, Sklearn, Linear Regression

Project Objective: By predicting the bike rental demand in advance from weather forcat, Bike rental company position the bike according to custmerrs demand resulting in increase in bike unitilization.

Quantifiable Results: We could predict the bike rentals Resulting in 95% accuracy and 135% increase in bike utilization

• Used Linear Regression to predict the number of bikes rented in the city of Seoul
• The data had quite a few categorical variables which were encoded for use in the model
• Encoded categorical variables to numeric using Sklearn due to the presence of mWe can Predict the Customer’s Response with 91.25% accuracy. any string columns
• Fit a multiple linear regression model with high prediction accuracy through iteration

---

Predicting customer subscription for term deposit

Skills used: Python, Numpy, Pandas, Matplotlib, Seaborn, SKlearn, Logistic Regression, SMOTE, RFE

Project Objective: Used logistic regression to predict customer subscriptions for there term deposit (Used real worked data from UCI, machine learning repository)

Quantifiable Results: Used Logistic regression classifier got 91.03 % of model accuracy which helps the bank to estimate their depositors in advance

• Used logistic regression to explore and make note of attributes information
• Use data cleaning principles to get rid of the unwanted data
• Explore data with groupby function and analyzing it
• Used Logistic regression classifier to improve on the data to make better prediction on whether the client will subcribe or not
• Explored a machine learning approach to predict by using Sklearn numeric

---

Identifying symptoms of orthopedic patients as normal or abnormal

Skills used: Python,Numpy, Pandas, Matplotlib, Seaborn, Sklearn, K Nearest Neighbours

Quantifiable Result: Used the K Nearest Neighbours algorithm to classify a patient’s condition as normal or abnormal based on various orthopedic parameters and got accuracy of 82.26 % for KNN and 82.26 % for NB

• Compared predictive performance by fitting a Naive Bayes model to the data
• Selected best model based on train and test performance
• Tuned hyperparameter best score was 0.8586
• The accuracy of the KNN with K = 17 is 82.26%
• The accuracy of the NB is 82.26%

K-Means clustering model for State authorities/Policy makers

Skills used: Python,Numpy, Pandas, Matplotlib, Seaborn, Sklearn, K-Means clustering

Project objctive: Based on the K-Means clustering model we could give useful insights to the authorities to group the people into respective catagories

Quantifiable Result: Based on the model all were grouped into 3 clusters, so k = 3, which got justified by Schilhoutee score = 0.5259

• Fitted the data to clustering model
• Calculated the number of clusters for the data
• Fitted the model with K-Means and Hierarchical clustering
• Compared the resluts for both methods
• Calculated the accuracy of the model which was 0.5259

---

Implementing Deep Neutral Network with Keras for handwritting classification and recognition

Skills Used: Matplotlib, numpy, Seaborn, Sklearn, Python, Neural Networking, Keras, Tensorflow

Project Objective: Implementing Deep Neural Network with Keras for handwritting classification and recognition

Quantifiable Results:

• Loss = 0.037 - 0.9909 = Final Model Eval = 99.0899%
• Load the data
• Initial data indagation
• Initial EDA
• Data Prepping
• Model contruction
• Model eval
• Model eval metrics
• Improved the model
• Improved the depth of model

---

Apple Stock Price Pridiction

Skills Used: Matplotlib, numpy, Seaborn, Sklearn, Python, LSTM, Neural Networking, Keras, Tensorflow

Project Objective:Stock price prediction Using Python, & Machin Learing (LSTM. Creat artificial neural network called Long Short Term Memory to predict the feuter price of stock:

• Initial data indagation
• Initial EDA
• Data Prepping
• Model contruction
• Model eval
• Model eval metrics
• Improved the model
• Improved the depth of model