Weather-Patterns-Analysis-and-Prediction.pdf
- 1. Weather Patterns Analysis and Prediction using machine learning This is a project report for the pattern analysis of weather dataset provided. We’ll look into some predictions made using machine learning models and key features of dataset. by Yash Saxena
- 2. Project Overview Objective Predict weather conditions (e.g., rain presence) Identify weather patterns and clusters Use machine learning algorithms for prediction Dataset Overview The Dataset shows a number of aspects of weather. Temperature, weather condition and rain presence are a few of such. It has 730 instances, 8 features and 1 Key column. Dataset contains weather data from 1st jan 2015 to 31st march 2017.
- 3. Exploratory Data Analysis (EDA) 1 Temperature Lowest: 12°C Highest:45°C Average:30.78 °C Frequently Occuring:35°C Occuring in 70 times. 2 Humidity Lowest:6% Maximum:100% Average:36.34% Frequently Occuring:31% Occuring 30 times 3 Pressure Lowest:994 hPa Highest:1026 hPa Average:1007.742 hPa Frequently occuring:1014 hPa Occuring times 4 Dew Point Humidity Lowest:1 °C Highest:28 °C Average:16.64 °C Frequently occuring:12 °C Occuring times
- 4. Methodology:Learning algorithms I used K-NN Classification The K-Nearest Neighbors algorithm calculates distances between data points using a chosen metric (e.g., Euclidean distance) and then assigns a class based on the majority class among the k-nearest neighbors. Here, k is the number of neighbours taken in consideration. This number is input by user. K-Means Clustering K-Means Clustering aims to partition data into k number of clusters by choosing k number of centroids and assigning membership to nearest points and then recompute their locations and the process goes on until there is no more recomputation of their locations.
- 5. Results: Model Performance K-NN Classification Using this (where k=3), the rain presence predicted on the 06th Jan 2019 was 0, indicating no rain. This output came from the three nearest neighbours which were the following--- K-Means Clustering The K-Means algorithm successfully identified distinct clusters within the weather dataset with the two centroids as 01-Jan- 2015 and 02-Jan-2015 DATE 14-Jan-2016 28-Jan-2017 29-Jan-2017 Distance 4.034 4.585 2.833
- 6. Insights and Learnings Trends Analysis uncovered that the lower the humidity was, the higher the visibility became. Insights While analyzing, it was uncovered that there is a steady increase of 0.88 degree celsius every year. This can raise a big concern regarding Global warming. With other natural phenomenons like El Niño, there is a good possibility in excessive increase in temperature. Impact Machine learning techniques proved effective in analyzing and predicting weather patterns, demonstrating the potential of data science and AI in advancing weather forecasting and improving decision-making.
- 7. Challenges and Recommendations 1 Challenges The biggest challenge was to use clustering in the 7 dimensional format, and assigning custom points for the K means clustering. 2 Recommendations Future projects could benefit from using more sophisticated machine learning models, such as deep learning algorithms, to capture complex weather patterns and improve predictive accuracy. 3 Recommendations With more variety in the dataset, the model accuracy and prediction could have been better.
- 8. Conclusion This project demonstrated the successful application of machine learning techniques for analyzing and predicting weather patterns. The results suggest that AI and Machine learning can become future of Data Analysis. The help of AI can be taken in other fields like Stock Market prediction, Generative AI, and many more.
- 9. References Google Sheets Libre Office Impress(for presentation) Gamma app (Template of ppt) Orange(Machine learning algorithms)