/Clustering-Geolocation-Data-Intelligently-in-Python

This is a python Coursera guided project successfully completed by me.

Primary LanguageJupyter Notebook

Clustering-Geolocation-Data-Intelligently-in-Python

This is Coursera Guided Project completed by me with the following learning objectives:-

  1. How to visualize and understand geographical data in an interactive way with Python.

  2. How the K-Means algorithm works, and some of the shortcomings it has.

  3. Density-based clustering approaches, and how to deal with any outliers they may classify.

Initially the project was completed by me on the Coursera's hands-on platform "Rhyme", but later I downloaded ht Jupyter Notebook and saved my progress.

Following python modules/functions have been used in the project:-

  1. matplotlib for plots and charts visualization of the outcomes.

  2. Pandas for storing and manipulating data.

  3. Numpy for its use in data-manipulation.

  4. hdbscan and DBSCAN for spatial-clusterings (hierarchichal).

  5. sklearn functionalities like Kmeans and silhouette_score with KneighboursClassifier.

  6. folium for maps and co-ordinates visualization.

The Project has been divided into 7-tasks:-

Task 1: An introduction to the problem, as well as basic exploratory data analysis and visualizations.

Task 2: Visualizing geographical data in a more meaningful and interactive way.

Task 3: Methods of evaluating the strength of a clustering algorithm.

Task 4: Theory behind K-Means, and how to use it for our problem.

Task 5: Introduction to density-based clustering approaches, and how to use DBSCAN.

Task 6: Introduction to HDBSCAN, to alleviate constraints of classical DBSCAN.

Task 7: A simple method to address outliers classified by density-based models.

At the end of this Project I found out that I need to work more on :-

  1. K-Means Algorithm.

  2. Density-based clustering approaches with HDBSCAN.

  3. A little bit of DataVisualization skills.