Analyzed and visualized earth surface temperature data provided by Berkeley Earth

1. Regression
   • Temperature prediction and visualization for a specific country, city, month and years using Linear Regression.
   • Temperature prediction for a specific year using multiple regression techniques like
     1. Linear regression
     2. Isotonic regression
     3. Bayesian Ridge
     4. RANSAC
     5. Gaussian Process
     6. SVR
2. Clustering
   • Classified earth surface temperature into 3 clusters of hot, mild and cold.
   • Generated a google map with showing the clusters for a specific year.
   • Model can predict temperature class of any given latitude and longitude for any year and month.
3. Visualization
   • Earth surface mean temperature rise curve from year 1750.
   • Divided a year in 4 seasons and visualized the average temperature of major cities in India.
   • Visualized earth surface average temperature using world map.

• Python Flask
• Spark
• Scikit
• AngularJs
• Google Maps
• Google Charts
• Plotly Js
• Matplotlib